DE60022237T2 - Bid processing method and system and recording medium for carrying out the method - Google Patents

Description

BACKGROUND THE INVENTION

Territory of invention

The The present invention relates to a method and a system for Carry out an image processing according to the preamble of claim 1 or claim 9. The present invention also refers to a computer-readable recording medium according to the preamble of the claim 17th

description of the prior art

We have proposed various image processing methods and systems for improving the diagnostic performance of a radiation image by performing frequency emphasis processing or dynamic range compression processing using a blur mask image signal with respect to a radiation image signal representative of the radiation image (hereinafter referred to as a "blur image signal"). For example, refer to Japanese Unexamined Patent Publication Nos. 55-163472, 557953, 3-222577, 10-75395 and 10-171983, for example, in the frequency emphasis processing, for example, a predetermined spatial frequency component becomes of an original image signal is emphasized by subtracting a blur image signal S _us from the original image signal S _orig and the remainder multiplied by a blur or emphasis coefficient β onto the original image signal Sa. ng is added. This is described by the following formula (1): S proc = S org + β × (S. org - p us ) (1) where S _{proc is} a frequency emphasized image signal, S _{org is} an original image signal, S _{us is} a blur image signal, and β is a highlight or emphasis coefficient.

In Japanese Unexamined Patent Publication No. JP 10-75395, there is also described a method for preventing the generation of an artifact in the frequency-emphasized image signal by adjusting the frequency response of the addition signal to be added to the original image signal. In these methods, a plurality of blur image signals different in frequency response, that is, sharpness, are prepared, differences are formed between each of the original image signal and a blur image signal to thereby receive a plurality of bandlimited signals having associated frequency components in the restricted frequency bands of the original image signal to transform the band-limited signals thus obtained into signals having desired values by using various transform functions, whereupon the addition signal is formed by adding up the band-limited signals. This is described, for example, by the following formula (2): S proc = S org + β (S. org ) × F usm (S org , P us 1, p us 2, .... S us N) F usm (S org , P us 1, p us 2, .... S us N) = f 1 (S org - p us 1) + f 1 (S us 1 - p us 2) + ... + f k (S us k - 1 - p us k) + ... + f N (s us N - 1 - S us N) (2) wherein S _proc is a processed image signal S _org, an original image signal, S _us k (k = 1 to N) is an unsharp image signal, f _k (k = 1 to N) is a transformation function, and β (S _org) a value determined on the basis of the original image signal emphasis coefficient is.

In addition, in Japanese Unexamined Patent Publication No. JP 10-171983, a method for preventing the occurrence of an artifact in the processed signal when performing both the frequency emphasis processing and the dynamic range compression processing is disclosed. In this method, a plurality of band-limited signals are formed as described above, on the basis of these band-limited signals, a high-frequency component signal representative of high-frequency components of the original picture signal and a low-frequency component signal representative of low-frequency components of the original picture signal are generated, and the frequency emphasis processing and the dynamic range Compression processing is performed by adding the high frequency component signal and the low frequency component signal to the original image signal. This is described, for example, by the following formula (3): S proc = S org + β (S. org ) × F usm (S org , P us 1, p us 2, ..... S us N) + D {p org - F drc (S org , P us 1, p us 2, ..... S us N)} F usm (S org , P us 1, p us 2, ..... S us N) {f ul (S org - p us 1) + f ul (S us 1 - p us 2) + ... + f uk (S us k - 1 - p us k) + ..... + f U.N (S us N - 1 - S us N) F drc (S org , P us 1, p us 2, ..... S us N) = f d1 (S org - p us 1) + f d2 (S us 1 - p us 2) + ... + f dk (S us k - 1 - p us k) + ... + f dN (S us N - 1 - S us N) (3) wherein S _proc is a processed image signal, S _org, an original image signal, S _us k (k = 1 to N) unsharp image signals, f _uk (k = 1 to N) transformation functions for obtaining the high frequency component signal, f _dk (k = 1 to N) is a transformation function for obtaining the low-frequency _{component signal} , β (S _org ) is an emphasis _coefficient , which is determined on the basis of the original _{image signal} , and D {S _org -F _drc (S _org , S _us 1, S _us 2, .. .. S _us N)} is a dynamic range compression coefficient determined on the basis of the low frequency _{component signal} , where D is a function of transforming D {S _org - F _drc (S _org , S _us 1, S _us 2, ... . S _us N)} is.

at the frequency emphasis processing and the dynamic range compression processing (which in the following representative be referred to as "transformation processing") can the frequency response of the to be added to the original image signal Addition signal can be adjusted by the definition of the transformation functions and the like for transforming the band-limited signals is changed. consequently let yourself by properly defining the transformation functions a processed Win picture signal, which has a desired frequency response, for example for the prevention of the emergence of an artifact is suitable. Indeed it is not easy to gain knowledge about how the transformation functions to define a desired Result. Therefore, in Japanese Unexamined Patent Publication No. JP 10-63838 proposed a method in which a processed Picture signal with a desired Frequency response is obtained in a simple manner that for a processed Picture signal a desired frequency response is set and parameters be determined to the transformation functions (this parameter hereinafter referred to as "transformation function definition parameter") to define the basis of the specified frequency response.

The Unsharp image signals, those in the above explained Transformation processing used are obtained by that by Filtering pixels of the original image signal at predetermined intervals and interpolating a corresponding number of pixels a thinning out of picture elements. As a filter processing was a processing for eliminating high frequency components from the original image signal by using a low-pass filter, in particular, a processing for calculating an average value or a weighted average of the pixel values in carried out the filter. In the filter processing with the purpose of multiple blur image signals as stated in Japanese Unexamined Patent Publication No. IP 10-75395 or the like, the blur image signals become gained by the fact that the Original image signal is filtered to produce a picture signal with less To get picture elements, the picture signal with the fewer picture elements is filtered and into the image signal with the fewer pixels, which was obtained by each filtering stage, picture elements be interpolated such that the Number of picture elements in the picture signal is equal to that of the original picture signal.

each Unsharp image signal is thus created on the basis of the original image signal by reading an original image with a predetermined reading density using an image reader and digitizing of the thus obtained image signal to a digital image signal, which reproduce an image with a predetermined pixel density can. It is known that frequency components below a certain, determined by the pixel density Frequency (a Nyquist rate) can then be played correctly, when a digitized image signal as a printed output element to reproduce. This means: as they are considering the picture quality, which is required during playback has been set is the Reading density, that is the pixel density, not constant.

In a radiation image reading and reproducing system, for example, in which a radiation image of a human body, which recorded on a stimulable phosphor sheet as a digital image read by scanning the phosphor sheet with a laser beam, The reading density or the pixel density varies depending on the size of the excitable Phosphor sheet and settles freely set by a user.

If Image signals of different pixel density or Nyquist frequency the same filter processing using the same low-pass filter and subsequently undergo the same interpolation differ the frequency responses the band - limited signals (in particular the frequency bands of the obtained band-limited signals) depending on the pixel density. This means that, for example when a pair of image signals of different pixel density obtained by reading an original image with different reading densities and frequency emphasis processing or dynamic range compression processing in terms of the image signals are made by using bandlimited Signals obtained on the basis of the same blur image signals the stressed frequency band or the compressed frequency band for the two Original image signals is different.

Around overcome this problem was in the Japanese unchecked Patent publication No. JP 10-6837 proposed a method in which blur image signals be obtained that from a List of filter coefficients depending on information about the Pixels density of the original image signal filter coefficients are selected and the original image signal using filters with the selected filter coefficients is filtered. This means: For example, if original image signals associated with read densities of 5 lines / mm and 6.7 lines / mm were read, with the help of the same low pass filter are filtered, the obtained two band-limited Signals from each other in the frequency band. However, at the in the above-mentioned Japanese Unexamined Patent Publication proposed method, the two original image signals from different low-pass filters filtered, and therefore can the two band-limited signals essentially the same frequency band exhibit. Accordingly, blur image signals same frequency bands independently be obtained from the pixel density of the original image signals, whereby band-limited signals of the same frequency response are possible and consistently a desired transformation processing throughout carried out may, for example, the frequency emphasis processing discussed above.

There however, the energy of one represented by a band limited signal bandlimited image, that is the peak value of the frequency response of a band-limited signal, dependent varies from the pixel density, the method according to Japanese unaudited Patent publication No. JP 10-63837 the response of the band-limited signals within the same frequency band is not independent of the pixel density although it is the frequency bands of the band-limited signals can match each other. So even if processing is done, to highlight a bandlimited signal in a particular frequency band, the frequency response of the emphasized band-limited signal varies depending on the pixel density.

In The image processing system described above sometimes becomes Object original image signal (an original image signal to be processed) inputted, which is a Picture corresponds to that in the resolution differs from images normally used by the image processing system are processed. (The resolution the images that are normally processed by the image processing system will be referred to as "reference resolution" below) in such a case, the transformation processing, for example the frequency emphasis processing, with respect to the object original image signal performed by using the transformation functions, the for Set original image signals that represent images with the reference resolution, it is to be feared that the Frequency response of the image, which is based on the object original image signal is reproduced from the image signal obtained is different from that of the Image based on the image represented by a reference resolution image Original image signal is reproduced. This problem can be solved overcome for example by that several Groups of transformation functions are created and corresponding the resolution of the image represented by the object original image signal is a group the transformation functions is used. However, this one has Approach the disadvantage that the Number of transformation functions too large and the handling of transformation functions too laborious becomes.

Although various formats such as JPEG, Gff, TIFF, and the like are known as formats for compressing an original image signal, a format has recently been proposed according to which an original image signal has been hierarchically decomposed into hierarchical data according to resolution and encodes the hierarchical data within each hierarchy level and be compressed. In this compression format, in particular, an original image signal is decomposed by a wavelet transform or the like into a plurality of hierarchical image signals each having a resolution of 1/2 ⁿ that of the original image signal, and the hierarchical image signals are encoded within the hierarchical sequence and compressed into a single file ,

• (1) Da das Bildsignal nicht blockweise verarbeitet wird, im Gegensatz zu dem beim konventionellen JPEG verwendeten DCT-Format, kommt es zu keinem Artefakt in Form einer Block-Verzerrung.
• (2) Da die Bildsignale hierarchisch kodiert werden, muß lediglich Information über notwendige Auflösungen nach dem Transfer der Bildsignale transferiert werden, was zu einem effizienten Bildtransfer beiträgt.
• (3) Da das Bildsignal in mehrere Auflösungen zerlegt wird, können verschiedene Bildverarbeitungen, beispielsweise eine Bildbetonungsverarbeitung, in relativ einfacher Weise durchgeführt werden.
• (4) Da durch eine Mehrfachauflösungsanalyse eine räumliche Zerlegung und eine frequenzmäßige Zerlegung gleichzeitig dwchgeführt werden können, kann eine orthogonale Transformation in einem breiten Bereich in der Niederfrequenzzone erfolgen, was in starkem Maß Einfluß auf die Kodierungseffizienz hat, wenn auch nur in einer schmalen Zone des Hochfrequenzbereichs. Selbst wenn also im Randbereich des Bilds Quantisierungsrauschen entsteht, läßt sich die räumliche Aufspreizung des Rauschens derart unterdrücken, daß man das Rauschen weniger wahrnimmt.

The compression format has the following characteristics:

• (1) Since the image signal is not processed block by block, unlike the conventional JPEG DCT format, there is no artifact in the form of block distortion.
• (2) Since the image signals are hierarchically coded, only information about necessary resolutions after the transfer of the image signals has to be transferred, which contributes to an efficient image transfer.
• (3) Since the image signal is decomposed into multiple resolutions, various image processing such as image enhancement processing can be performed relatively easily.
• (4) Since spatial decomposition and frequency decomposition can be simultaneously performed by a multi-resolution analysis, orthogonal transformation can be performed in a wide range in the low frequency zone, which has a great influence on the coding efficiency, albeit only in a narrow zone of the high frequency range. Even if quantization noise is generated in the edge region of the image, the spatial spread of the noise can be suppressed in such a way that one perceives the noise less.

In addition, were various file formats are proposed, for example those of Eastman Kodak proposed F1ashPix file in which data is different Types can be stored in a single file. The Hierarchical image signals discussed above may also be used in such a manner FlashPix standard file.

By decomposing an original image signal into multiple resolutions, it is possible to construct an original image signal by plural hierarchical image signals each having a resolution of 1/2 ⁿ times the original image signal. This makes it easy to reconstruct an image based on a part of the hierarchical image signals selected according to the image quality required by the output system. That is, when high image quality is required in reproduction in, for example, a printer, by reconstructing the image signal based on hierarchical image signals up to the highest resolution, a high quality image signal corresponding to that of the original image can be reproduced. For example, unlike, for example, a CRT that can not reproduce a picture with as high a quality as a printer, an image signal can reproduce an image suitable for the CRT, but the quality in resolution is not as high as that of the CRT Original image by reconstructing the image signal on the basis of the hierarchical signals that do not include the highest-resolution hierarchical image signal, and enlarging or contracting the image signal as needed.

There however, the hierarchical image signals each represent one image, which is a lower resolution (Hereinafter, these hierarchical image signals are referred to as "lower hierarchical image signals ") has as the original image signal, different in frequency response are from the original image signal, so when the transformation function for the Original image signal as they are in the frequency emphasis processing the lower hierarchical image signals are used to fear the existence Picture signal can be obtained, which differs in the frequency response from a signal generated by frequency emphasis processing of Original image signal is obtained. This can be avoided by Create a number of transformation functions accordingly different resolutions and using transformation functions according to the resolution of the processed one Image signal. However, this approach has the disadvantage that the number becomes too large to be handled transformation functions and the handling of the functions becomes too cumbersome. The problem emerges not only when frequency emphasis processing of those Perform image signals which is obtained by decomposing an original image signal into multiple resolutions but also when frequency emphasis processing a picture signal with the destination, a picture with a lower one resolution as the original image signal to reproduce.

Farther is for example one with intervals of 10 lines / mm by doubling and interpolating one from one stimulable phosphor sheet with a reading density of 5 lines / mm read image signal in its sharpness weaker than a radiation image, which with intervals of 10 lines / mm reproduced on the basis of an original picture signal which was of a stimulable phosphor sheet with a reading density of 10 lines / mm, although the picture sizes are are the same. The reason for that is that high frequency components of the original picture signal from the frequency response of the filter used to change the pixel density of the original image signal, that is, to contract and interpolate of the original image signal and the frequency response of the filter for doubling of the contracted original image signal are attenuated, and because the former Image differs from the latter image in the frequency response.

In addition, if a low-resolution image based on a hierarchical image signal too which is obtained by decomposing an original image signal of an original image into multiple resolutions, the obtained image becomes inferior in sharpness to the original image, depending on the wavelet transform functions used in wavelet transformation. In addition, if a low-resolution image is made the same size as the original image, the high-frequency components of the original image are weakened depending on the frequency response of the filter for enlargement and interpolation, and the obtained image is inferior in sharpness to the original image. The same problem also occurs when the original image is to be expanded or contracted to a desired size. Thus, when the above-mentioned frequency emphasis processing is performed on an image signal which is a less sharp image than the original image, the extracted image differs in expression from an image obtained by frequency emphasis processing of the original image signal.

According to the preamble of each of claims 1, 9 and 17 shows the EP 0 756 247 a method, a system and a computer-readable medium for decomposing an image into a plurality of images of different frequency bands, performing frequency emphasis processing of the respective images so as to emphasize the images in predetermined frequency bands, and performing inverse transformation processing on the images having emphasis processing was performed to reproduce an image. However, the processing performed in the prior art does not take into account the fact that the frequency response of the image varies depending on the resolution.

in the In view of the above representations and explanations, it is the main objective of the invention, a method and system for image processing specify a desired transformation processing in a constant manner, For example, a processing for highlighting a special one Frequency component can, to be independent from the pixel density of the original image signal gives the same result achieve.

Reached This is achieved by the features of claim 1 and claim 9. A computer readable medium is defined by claim 17. Preferred embodiments are in the dependent claims Are defined.

The present invention provides a method and system for image processing, which can perform frequency emphasis processing in a constant manner, which leads to the same result regardless of the resolution of the leads to processed image, without the need for a large number of transformation functions save.

In addition, creates the present invention provides a method and system for image processing, which constantly provides frequency emphasis processing in perform the way that can happen the same result independently from the resolution of the image to be processed, without the need for a size Number of transformation functions to save.

In addition, creates the present invention provides a method and system for image processing, which can perform frequency emphasis processing of an image which in terms of pixel density and / or scale from an original image therefore, the frequency response of the obtained image is equivalent to that of the original picture is.

The The present invention also provides a computer readable Recording medium loaded with a program which cause a computer to an image processing according to the inventive method perform.

The "transformation processing" includes, for example a frequency emphasis represented by the formula (2) for emphasizing a special frequency component, and dynamic range compression processing according to the formula (3) for reducing the contrast of the high-density region and / or low density region, thereby reducing the dynamic range narrowing, which is the difference between maximum Density and minimum density of the original image.

The "pixel density of the original image" can either be entered by the operator or automatically obtained when the original image signal is edited, in which case information about the pixel density is pre-added to the original image signal. "Pixel density" can be, for example, a value of resolution (e.g. dpi ") representing the relation between the size of the original image and the sampling intervals for obtaining the original image signal as well as the reading density with which a radiation image recorded on a stimulable phosphor sheet is read.

Farther the term "transformation function definition parameter for the Transformation functions are based on the pixel density of the original image signal determines, for example, that the Parameters are obtained by resolving relational expressions of explained above several Intermediate image signals, a desired frequency response and the transformation function definition parameter in the form of simultaneous equations using the plurality of intermediate image signals and the desired frequency response as known values and the transformation function definition parameter as a variable, or around the transformation function definition parameters by gradually changing the Parameters while observing the on the basis of the processed image signal reproduced image. As transformation functions can Various functions are used, such as linear Functions, non-linear functions, constant and the like.

It it is preferred that the several intermediate image signals that are in the frequency band from each other be created by the original image signal of a Filter processing is performed using filters, their filter coefficients based on the pixel density (= Resolution) of the original image to thereby obtain a plurality of blur image signals to create, which differ in the frequency response, a plurality of band-limited signals is created, the signals in the individual frequency bands represent the original image signal, based on the blur image signals and the original image signal, and the band-limited signals are used as intermediate image signals.

The "blur image signals" are image signals, which is equivalent to the original image signal in the number of pixels are, but represent pictures, the lower sharpness own than that represented by the original image signal Image. The blur image signals are made by thinning of picture elements by adding picture elements of the original picture signal in are filtered out at predetermined intervals, further picture elements thinning out by filtering out pixels of the thinned image signal, these procedures are repeated, and in the image signal with less Pixels obtained by each filter stage, pels such be interpolated that the Number of picture elements in the picture signal of those in the original picture signal like. The filtering is carried out using filters with filter coefficients, which are determined based on the pixel densities, such as this in detail in the above-mentioned Japanese Unexamined Patent Publication No. JP 10-63837.

The "several band-limited signals which the image signals in the respective Frequency bands of the For example, original images may represent be generated by differences between blur image signals in adjacent frequency bands, Differences between the original image signal and the respective blur image signals or differences between any other combinations of the blur image signals be taken.

Prefers The bandlimited signal generating means includes a Unsharp image signal generation means, the multiple blur image signals created, which differ in the frequency response by the original image signal is filtered with filters, their filter coefficients based on the pixel density (Resolution) of the original image, and several band-limited signals, which signals in the respective frequency bands of the original image signal represent, on the basis of the blur image signals and the original image signal.

According to the procedure and the system of the present invention, because the transformation function definition parameters based on the resolution (Pixel density) of the original image and the transformation processing according to the transformation functions which is determined by the transformation function definition parameters are defined, the parameters are set such that the in Frequency band transformed various band-limited signals are transformed into image signals that have the same frequency response have, independent from the resolution (Pixel density) of the original image, whereby the transformation processing done in the way can be that one processed image signal can reproduce an image, which is a has constant frequency response, without detrimental influence through the resolution (Pixel density) of the original image.

The "transformation processing" includes, for example, a frequency emphasis processing represented by the formula (2) for emphasizing a specific frequency component, and a dynamic range compression processing according to the formula (3) for decreasing the contrast of the high density range and / or low density range, thereby narrowing the dynamic range, which is the difference between the maximum and minimum density of the original image.

The "dissolution of the Object original image signal "can either entered by the operator or automatically at the processing of the original image signal are obtained. In the latter Case, the original image signal beforehand information about the resolution attached. The "resolution" can, for example a resolution value (expressed for example in dpi), which shows the relation between the size of the original picture and represents the sample intervals, where the original image signal is obtained, also the Reading density, with which one on a stimulable phosphor sheet recorded radiation image is read.

The "blur image signals" are image signals, which is equivalent to the original image signal in the number of pixels are, though for Images that are less sharp than those represented by the original image signal Have picture. The blur image signals are created by thinning of pixels by filtering pixels of the original image signal at predetermined intervals, by further thinning out picture elements by filtering pixels of the thinned image signal, by repeating of these procedures and by interpolating picture elements into the less picture element having image signal passing through each Filtering was obtained, such that the number of pixels in the image signal is equal to that of the original image signal. The Filter processing is performed using filters whose Filter coefficients determined based on the pixel densities were, as described in detail in the above-mentioned Japanese unaudited Patent publication No. 10 (1998) -63837.

The "several band limited signals which signals in the respective frequency bands of the Original image signal represent "can, for example be obtained by the Differences between blur image signals in adjacent frequency bands, Differences between the original image signal and the respective blur image signals or differences between any other combinations of Unsharp image signals be taken.

According to the procedure and the system according to the invention, because the transformation functions for transforming the intermediate image signals obtained from an object original image signal for a Image with a different resolution as the reference resolution, by correcting the reference transformation functions according to the resolution of the Object original image signal to be obtained, the transformation functions be set up so that the Intermediate image signals are transformed into transformed image signals, which independently from the pixel density of the original image signal has the same frequency response possess, whereby the transformation are carried out in such a way can this be processed image signal reproducing an image with constant Frequency response without interference from the resolution of the original image signal allows.

There Furthermore the transformation functions for Original image signals with resolutions, which differs from the reference resolution be obtained by correcting the reference transformation functions, it is not necessary to have a plurality of transformation functions for different To provide resolutions which eliminates difficulties in handling a variety of transformation functions.

The "transform functions set up according to the frequency bands of the respective band-limited signals" have a number of six when the intermediate-frequency signals are in six frequency bands.The transformation functions may be constants, the expression "the respective reference transform functions for the band-limited signals in the Frequency bands not above the frequency band corresponding to the resolution of the original image represented by the object original image signal "means, for example, if the resolution of the original image represented by the object original image signal is 1/2 ^k times the reference resolution, the reference Transformation functions for the k-th intermediate image signal (band-limited signal) count from the high-frequency side and those of the frequency bands below the k-th intermediate image signals. For example, when six inter-image signals are obtained from the reference original image signal and six transformation functions (f ₁ to f ₆ , where f _{1 represents} the highest frequency band) for the respective intermediate image signals with the resolution of the original image represented by the object original image signal is 1/2 are set to those of the reference original image, the transformation functions f ₂ to f ₆ "are the respective reference transformation functions for the intermediate image signals in frequency bands not above the frequency band, which is the resolution of the original represented by the object original image signal picture "are.

The "blur image signals" are image signals equivalent to the original image signal in the number of pixels, but representing images having a lower sharpness than the image corresponding to the original image signal The blur image signals are obtained by thinning pixels by filtering pixels of the original image signal at predetermined intervals by further thinning pixels by filtering pixels of the thinned image signal, repeating these procedures to obtain a plurality of low-resolution image signals representative of images whose resolutions are 1/2 ^{n of} that of the image represented by the reference original image signal, and by interpolating picture elements into the low-resolution picture signal having fewer picture elements obtained in each filter stage, such that the number of picture elements in the picture signal is equal to that of the original picture signal.

The "several band limited signals representing the signals of the respective frequency bands of the Original image signal represent "can, for example be obtained by the Differences between blur image signals in adjacent frequency bands, Differences between the original image signal and the respective blur image signals or differences between other combinations of the blur image signals be taken.

Farther It is preferred that information about the resolution of the original image, represented is obtained by the object original image signal, the step setting the transformation functions to transform the intermediate image signals obtained from the object original image signal will be executed on the basis of this information.

The resolution information can be overridden by the operator a keyboard can be typed or entered by the operator by selecting a value from several displayed on a control screen resolutions be determined. Furthermore can the resolution information attached to the object original image system and from the image processing system be recognized. The resolution information can be entered in any way, as long as the system resolution of the object represented by the object original image Can recognize original image.

In the image processing method of the present invention, when the original image represented and to be processed by the object original image signal has a lower resolution than the reference resolution, the transform functions for transforming the inter-image signals (= band-limited signals) obtained from the object original image signal are equalized the respective reference transformation functions for the intermediate image signals in the frequency bands not above the frequency band which corresponds to the resolution of the original image represented by the object original image signal. When the resolution of the original image 1/2 ⁿ corresponding to the object original image signal corresponds to the resolution of the reference original image represented by the reference original image signal, the peak frequencies of the intermediate image signals obtained from the object original image signal are the same as those of a part of the reference intermediate image signals (obtained from the reference original image signal) which are not higher in frequency bands than the frequency band corresponding to the resolution of the original image represented by the object original image signal. Accordingly, when the intermediate image signals obtained from the object original image signal are acquired on the basis of the reference transformation functions for that portion of the reference intermediate image signals whose frequency bands are not above the frequency band corresponding to the resolution of the original image corresponding to the object original image signal the obtained transformed image signals in their frequency response substantially the same as the transformed reference image signals. The processed image signal may reproduce an image having substantially the same frequency response as an image reproduced on the basis of the processed image signal obtained from the reference original image signal. Consequently, by the image processing method according to the invention, a processed image signal can reproduce an image which is substantially constant in frequency response regardless of the resolution of the object original image signal.

There Furthermore the transformation functions for obtaining the processed image signal be used from the reference original image signal, the processed It is to obtain image signal from the object original image signal not necessary, a plurality of transformation functions for different resolutions to provide, which simplifies the system in its construction can be and handling a variety of transformation functions can be designed without difficulty.

at the image processing method according to the invention become from the object original image signal the Reference image signals of low resolution from the reference original image signal, a reference image signal of low resolution for an image whose resolution is the Target resolution the next comes as object original image signal to obtain a contraction-processed Image signal, and obtained a plurality of image signals of low resolution. As a result, the peak frequencies of the band-limited signals are lower Resolution, which are obtained from the object original image signal, the same as those of a part of the band-limited reference signals lower Resolution, in frequency bands are not higher are the frequency band that corresponds to the resolution that is the target resolution at next comes. So if the band-limited signals of low resolution, the were obtained from the object original image signal on the basis of the reference transformation functions for the part of the band-limited Transform low-resolution reference image signals, in frequency bands are not above the frequency band whose resolution the Target resolution on next comes, then have the obtained transformed image signals one Frequency response that is essentially the same as the one the transformed reference image signals. The contraction-processed Image signal obtained by enlarging or contracting the image signal processed intermediate image signal obtained from the transformed image signals in such a way that the resolution represented by the processed intermediate image signals Image resembles the target resolution, can reproduce an image which is essentially the same Frequency response has as an image based on the processed Reference picture signal is reproduced. Thus, regardless of the resolution of the image to be reproduced, the image processing method according to the tenth Aspect of the invention provide a contraction processed image signal, which can reproduce an image which has a constant frequency response Has.

There Furthermore the transformation functions for obtaining the processed reference image signal thereto be used to obtain the contraction-processed image signal, it is not necessary to have a plurality of transformation functions for different resolutions to provide, which simplify the system in its structure leaves and difficulties avoided when handling multiple transformation functions can be.

SHORT DESCRIPTION THE DRAWINGS

1 Fig. 10 is a schematic block diagram showing the configuration of an image processing system according to a first embodiment of the invention;

2 FIG. 12 is a schematic block diagram of an embodiment of the blur image signal generating means used in the in FIG 1 shown image processing system is used

3A and 3B are views of examples of the filter used in filter processing,

4 shows in more detail the process of image signal generation for a low-resolution image,

5 FIG. 14 is a view of an example of the filter used in the interpolation enlargement processing; FIG.

6 Fig. 10 is a view of a system for performing frequency emphasis as non-linear processing;

7 shows the frequency response of band-limited signals obtained by performing the filter processing with the in 3A shown one-dimensional filter with respect to an original image signal with a pixel density of 10 lines / mm,

8th shows the frequency response of band limited signals obtained by performing the filter processing with the in 3B shown one-dimensional filter with respect to an original image signal with a pixel density of 6.7 lines / mm,

9 shows the frequency response of band-limited signals obtained by performing the filter processing with the in 3A shown one-dimensional filter with respect to an original image signal with a pixel density of 5 lines / mm,

10 Fig. 10 is a schematic block diagram of the construction of the transformation function defining means;

11 FIG. 16 is a view of the frequency response of the processed image signal obtained by processing the band-limited signals obtained from an original image signal at 10 lines / mm by the image processing system of the first embodiment; FIG.

12 FIG. 16 is a view of the frequency response of the processed image signal obtained by processing the band-limited signals obtained from an original image signal at 6.7 lines / mm by the image processing system of the first embodiment; FIG.

13 Fig. 14 is a view of the frequency response of the processed image signal obtained by processing the band-limited signals obtained from an original image signal at 5 lines / mm by the image processing system of the first embodiment;

14 FIG. 14 is a view of an example of the user interface for determining the frequency response; FIG.

15 Fig. 10 is a flow chart for illustrating the operation of the image processing system of the first embodiment;

16 Fig. 15 is a view of a system for performing frequency emphasis processing and dynamic range compression processing as non-linear processing in an image processing system according to a second embodiment of the invention;

17 FIG. 14 is a view of an example of the transformation functions used in the first transformation part of the second embodiment; FIG.

18 FIG. 14 is a view of another example of the transformation functions used in the first transformation part of the second embodiment; FIG.

19 FIG. 14 is a view of an example of the transformation functions used in the second transformation part of the second embodiment; FIG.

20 _{FIG. 14} is a view of the frequency response of the processed image signal S _proc obtained by processing a 10 line / mm original image signal in the image processing system of the second embodiment; _FIG .

21 _{FIG. 14} is a view of the frequency response of the processed image signal S _proc obtained by processing an original image signal of 6.7 lines / mm in the image processing system of the second embodiment; _FIG .

22 _{Fig. 12} is a view of the frequency response of the processed image signal S _proc obtained by processing a 5 line / mm original image signal in the image processing system of the second embodiment;

23 Fig. 10 is a schematic block diagram showing the configuration of an image processing system according to a third embodiment of the invention;

24 is a view of the frequency response of band-limited signals, which were obtained from an original image signal S _org with a pixel density of 10 lines / mm,

25 _{FIG. 12} is a view of the frequency response of the processed image signal S _proc obtained by performing the frequency _emphasis processing using the transformation functions shown in Table 3 with respect to the original image signal S _org at 10 lines / mm (solid line) and the frequency response of the processed image signal S _proc was obtained by performing the frequency _{emphasis processing} using the transformation functions shown in Table 3 with respect to the original image signal S _org at 300 dpi (dashed line),

26 FIG. 16 is a view of the frequency response of the band-limited signals obtained from a 300 dpi original image signal; FIG.

27 FIG. 12 is a view of the frequency response of the processed image signal S _proc obtained by Verar processing of an original image signal at 300 dpi by the image processing according to the third embodiment of the invention,

28 Fig. 10 is a flowchart for illustrating the operation of the image processing system of the third embodiment;

29A and 29B FIG. 4 are views for illustrating different ways of selecting a reference pixel density; FIG.

30 is a view of the relation between the number of band-limited signals and the resolution,

31 FIG. 16 is a view of the frequency response of the band-limited signals obtained from a 200 dpi original image signal; FIG.

32 shows the frequency response of the processed image signal S _proc obtained by processing the original image signal S _org at 200 dpi using the reference transformation functions (dashed line) and the processed image signal S _proc obtained by processing the reference original image signal S _org 0 with the reference Pixel density using the reference transformation functions (solid line),

33 is a view of the relation between the answer X1 [Fi] and the function X1 [i],

34 is a view of the relation between the functions X1 [i] and Z1 [i],

35 FIG. 14 is a view of a system for performing frequency emphasis processing and dynamic range compression processing as transformation processing in an image processing system according to a fifth embodiment of the invention; FIG.

36 FIG. 14 is a view of an example of the transformation functions used in the first transformation part of the fifth embodiment; FIG.

37 FIG. 14 is a view of another example of the transformation functions used in the first transformation part of the fifth embodiment; FIG.

38 FIG. 14 is a view of an example of the transformation functions used in the second transformation part of the fifth embodiment; FIG.

39A FIG. 12 is a view of the frequency response of the processed reference image signal obtained by transforming the band-limited signals from the reference image signal (with a pixel density of 5 lines / mm) using the reference transformation functions set to amplify the relatively low frequency components, and FIG Frequency response of the processed image signal obtained by transforming the band-limited signals from the original image signal by using the transform functions obtained by correcting the reference transform functions according to the pixel density (4 lines / mm) of the original image signal such that the frequency response of the processed image signal is that of the processed one Equal to reference picture signal in frequency bands not lower than 1/2 of the Nyquist frequency of the original picture signal to be processed,

39B FIG. 12 is a view of the frequency response of the processed reference picture signal obtained by transforming the band-limited signals from the reference picture signal (with a picture element density of 5 lines / mm) using the reference transform functions set so that the relatively high frequency components are amplified; Frequency response of the processed image signal obtained by transforming the band-limited signals from the original image signal by using the transform functions obtained by correcting the reference transform functions according to the pixel density (4 lines / mm) of the original image signal such that the frequency response of the processed image signal is that of the processed one Equal to reference picture signal in frequency bands not lower than 1/2 of the Nyquist frequency of the original picture signal to be processed,

40 Fig. 10 is a schematic block diagram of the configuration of an image processing system according to a sixth embodiment of the invention;

41A to 41C are views for illustrating the encoding of the hierarchical image signals,

42 FIG. 12 is a schematic block diagram of the configuration of the blur image signal generating device used in the in 40 shown image processing system is used

43 a view of an example of the filter to be used for the filter processing,

44 shows in detail the image signal generating process for generating a low-resolution image signal,

45 FIG. 14 is a view of an example of the filter used in the interpolation enlargement processing; FIG.

46 FIG. 14 is a view of a system for performing frequency emphasis processing; FIG.

47 shows the frequency response of band-limited signals obtained from the reference original image signal,

48 shows the frequency response of band-limited signals obtained from a low-frequency original picture signal having a resolution of 1/2 that of the reference original picture signal;

49 shows the frequency response of band-limited signals obtained from a low-frequency original picture signal having a resolution of 1/4 that of the reference original picture signal,

50 shows the frequency response of band-limited signals obtained from a low-frequency original picture signal having a resolution of 1/8 that of the reference original picture signal;

51 shows the frequency response of band-limited signals obtained from a low-frequency original picture signal having a resolution of 1/16 that of the reference original picture signal,

52 Fig. 10 is a flow chart illustrating the operation in the image processing system according to a sixth embodiment of the invention;

53 _FIG . 12 is a view of the correspondence between the blur image _signals obtained from the reference original image signal S _org and those obtained from a low resolution object image signal; FIG.

54 _FIG . 12 is a view of the correspondence between the blur image _signals obtained from the reference original image signal S _org and those obtained from a low resolution object image signal; FIG.

55 Fig. 10 is a view of correspondence between the band-stepped signal transforming functions obtained from the reference original image signal S _org and those for the band-limited signals obtained from a low-resolution object image signal;

56 Fig. 10 is a block diagram of a blur image signal generating means in the image processing system according to a seventh embodiment of the invention;

57 Fig. 12 is a view of the correspondence between the band-limited signals obtained from the reference original image signal S _org and those obtained from the object original image signals;

58 Fig. 12 is a view of the correspondence between the transform functions for the band-limited signals obtained from the reference original image signal S _org and those for the band-limited signals obtained from the object original image signals;

59 Fig. 10 is a schematic block diagram illustrating the configuration of an image processing system according to a seventh embodiment of the invention;

60 FIG. 4 is a view of the frequency response of the pixel density transformation filter; FIG.

61 is a view of the frequency response of the interpolation filter,

62 _{FIG. 16} is a view of the frequency response of the processed image signal S _proc obtained by performing the frequency _emphasis processing using the transform functions f _k for frequency emphasis of the basic original image signal S _org 'with respect to the 5-line transformed original image signal S _org 5 and by doubling the obtained processed intermediate _{image signal} S _proc 'by using the interpolation filter with the filter characteristic F2,

63 Fig. 12 is a view of the frequency response of the band-limited signals obtained from the basic original image signal S _org ';

64 Fig. 10 is a flow chart illustrating the operation of the image processing system according to the seventh embodiment of the invention;

65 Fig. 10 is a schematic block diagram illustrating the configuration of an image processing system according to an eighth embodiment of the invention;

66A to 66C are views for illustrating the coding of the hierarchical image signals,

67 Fig. 12 is a view of the frequency response of the low-resolution image signals restored to a resolution equal to 1/2 that of the original original image signal S _org 'and restored to a resolution of 1/4 that of the original original image signal S _org ' when this basic original image _signal S _org 'has been decomposed using wavelet transform functions with first filter coefficients,

68 shows the same thing in the case where the basic original image _signal S _org 'has been decomposed using wavelet transform functions with second filter coefficients,

69 Fig. 11 is a view of an example of table data storing a spatial frequency characteristic corresponding to the enlargement characteristic or the reduction characteristic, and

70A and 70B are views of the parameters used in a Gaussian function representing the spatial characteristic, which in turn represents the magnification or reduction characteristic.

DESCRIPTION THE PREFERRED EMBODIMENTS

in the The following will be an image processing system according to a first embodiment of the invention. The image processing system is used for Carry out a frequency emphasis processing (hereinafter: frequency emphasis or highlight) by using blur image signals with respect to a Original image signal, which was obtained by reading a Radiation image of a human body on a stimulable Phosphor sheet was recorded so that a suitable for a diagnosis Image can be obtained. A based on the processed Image signal reproduced image is primarily based on a photographic Film recorded and used in the diagnosis.

In 1 contains an image processing system 1 According to a first embodiment of the invention, a blur image signal generator 2 which generates a plurality of blur image signals from an original image signal S _org input from an image reader or the like and has a predetermined pixel density, and a non-linear processing means 3 which performs frequency emphasis to emphasize a particular frequency and which produces a processed image signal S _proc . The image processing system 1 also includes a transformation function defining means 4 and a pixel density information input device 5 , The transformation function defining means 4 is a device that defines transformation functions involving the non-linear processing device 3 in the transformation processing, for example, by setting parameters such as the slope of the functions and the degree of non-linearity on the basis of the pixel density information M obtained from the pixel density information input device 5 was entered to define the functions based on these parameters. The picture element density information ons input device 5 is a means for obtaining pixel density information M via the original image signal S _org . The pixel density information M may be entered as a value by the operator via a keyboard, or may be selected by the operator from multiple pixel densities displayed on a control screen. In addition, the pixel density information M in the reader may be added to the original image signal S _org or may be provided by the image processing system 1 be detected. The pixel density information M may be input in any manner as long as the transformation function defining means 4 can recognize the pixel density of the original image signal S _org .

First, the generation of the blur image signals will be explained below. As in 2 is shown, contains the blur image signal generating means 2 a first to nth filter device 10 , The first filter device 10 performs filter processing on the original image signal S _org in the x and y directions, and generates a low resolution image signal B1 lower than that of the original image signal S _org . The second filter device 10 performs similar filtering processing on the thus obtained low-resolution picture signal B1 and generates a low-resolution picture signal B2 whose resolution is lower than that of the low-resolution picture signal B1, and the third filter means 10 performs similar filtering processing on the thus obtained low-resolution image signal, B2, and generates a low-resolution image signal, B3, whose resolution is lower than that of the low-resolution image signal, B2. The nth filter device accordingly leads 10 performs similar filtering processing on the low-resolution image signal Bn-1 and produces a low-resolution image signal Bn whose resolution is lower than that of the low-resolution image signal Bn-1. The blur image signal generator 2 also contains a first to nth interpolation device 11 , The interpolation device 11 performs interpolation / enlargement processing on the low resolution image signals B1 to Bn obtained in the respective stages of filtering to produce a plurality of unsharp image signals S _us 1 to S _us N having different sharpness.

at this particular embodiment be for the filter processing uses filters that are essentially one one-dimensional Gaussian distribution correspond. This means: Filter coefficients are determined according to the following Formula (4) with respect a Gaussian signal certainly:

This is because the Gaussian signal is a good location both in real space and in frequency space. For example, the 7 × 1 one-dimensional filters are configured as shown in FIG 3A is shown when σ = 1 in the formula (4).

The filtering is performed with respect to the original image signal Sorg or with respect to the low resolution image signals for every other picture element in the 4 performed manner shown. When such filtering is performed in both the x and y directions, the number of picture elements in the low resolution picture signal reduces to 1/4 of the preceding picture signal (for example, the original picture signal S _org in the case of the low resolution picture signal, B1 , and the low resolution image signal, B1, in the case of the low resolution image signal, B2). That is, the number of picture elements in each of the low-resolution picture signals B _k (k stands for 1 to n) is 1/2 ^{2k of} the original picture signal S _org .

The interpolation magnification processing to be performed on the low-resolution image signals Bk thus obtained will be explained below. Although various methods of interpolation may be used, such as the B-spline method, a Gaussian signal is used for interpolation in this particular embodiment because Gaussian-based low pass filters are used for filter processing. In particular, an approximation σ = 2 ^{k-1 is used} in the following formula (5):

When interpolating the low-resolution image signal, B1, because k = 1σ = 1, in this case, a one-dimensional filter of size 5 × 1, as shown in FIG 5 is shown used in the interpolation. In this interpolation, a picture element having the value 0 for every other picture element is inserted into the picture signal B1 of low resolution, so that this picture signal B1 is expanded to a size corresponding to that of the original picture, and the interpolated low-resolution picture signal B1 becomes less Use of the in 5 filtered one-dimensional filter shown.

This interpolation enlargement processing is performed on all the low-resolution image signals B1 to Bn. When a low-resolution image signal Bk is interpolated, a filter 3 × 2 ^k -1 in length is prepared in accordance with the formula (5), and 2 ^k -1 0-pixel pixels are inserted between each pair of adjacent pixels Image signal of low resolution, Bk, is expanded to a size corresponding to that of the original image signal. Then, the interpolated low-resolution image signal Bk is filtered by means of a filter having a length of 3 × 2 ^k -1.

The frequency emphasis was generally based on the assumption that the pixel density with which the original image S _{org is} obtained is constant, and the processed image signal S _{proc was} generated without taking into account the pixel density. However, in a practical image processing system, various image signals are input, and all the image signals do not have the same pixel density. For example, in the radiation image processing system of this embodiment, the pixel density differs depending on the size of the stimulable phosphor sheet, is 5 lines / mm at half size, 6.7 lines / mm at quarter size and 10 lines / mm at sixth size. In addition, the user can freely adjust the pixel density.

When image signals having different pixel density and corresponding to the Nyquist frequency are subjected to same low-pass filtering and interpolation processing, the band-limited signals obtained from the unsharp signals S _usk in the manner to be described later become different in frequency response depending on the pixel density. For example, when band-limited signals are generated from blur image signals S _usk , which are filtered by a one-dimensional filter according to FIG 3A with respect to original picture signals read at 10 lines / mm, 5 lines / mm and 6.7 lines / mm, the band-limited signals from the original picture signal of 6.7 lines / mm in the frequency bands are clearly different from those of Figs 10 line / mm and 5 line / mm original picture signals, although the latter band-limited signals in the frequency bands are the same, since the 5-line / mm original picture signal is half that of 10-lines / mm in Nyquist frequency. In this embodiment, the filtering is performed using the in 3A 10 filters / mm and 5 lines / mm original picture signals using the filter shown in FIG 3B of the original picture signal of 6.7 lines / mm so that the frequency bands of the obtained band-limited signals become substantially equal regardless of the picture element density of the original picture signal. The blur image signal generator 2 switches the filters on the basis of the pixel density information M transmitted via the pixel density information input device 5 is entered.

The non-linear processing to be performed on the blur image signals S _usk (k stands for 1 to n) will be described below. 6 shows a system for performing frequency emphasis (an example of the non-linear processing) together with the blur image signal generating means 2 , As in 6 As shown, the unsharp image signals S _us k from the original image signal S _org formed, and are of respective subtractors 21 Differences between the original image signal S _org and the blur image signals S _us k formed, whereby band-limited signals (S _org - S _us 1, S _us 1 - S _us 2 and the like) arise, which are components in limited frequency bands of the original image signal S _org ,

The band-limited signals are provided with different transformation functions f ₁ to f _N in respective transformation circuits 22 transformed to predetermined amplitudes, and the transformed band-limited signals are added and then by an operator 23 is added to the original image signal S _org according to the formula (2) given below, thereby producing a processed image signal S _proc in which a desired frequency component is raised to a desired level. S proc = S org + β (S. org ) × F usm (S org , P us 1, p us 2, ..... S us N) F usm (S org , P us 1, p us 2, ..... S us N) = f 1 (S org - p us 1) + f 2 (S us 1 - p us 2) + ... + f k (S us k - 1 - p us k) + ... + f N (S us N - 1 - S us N) (2) wherein S _proc the processed image signal, in which the high frequency components are emphasized S _org, an original image signal, S _us k (k = 1 to N) unsharp image signals, f _k (k = 1 to N) is a transformation function, and β (S _org ) is an emphasis coefficient determined on the basis of the original image signal.

As transformation functions f _k are non-linear functions according to the following formula (6) ver applies: f (S in ) = S out = S in × Y × {exp (X / S in ) - 1} / {exp (X / S in ) + 1} (6) wherein S _in an input signal, S _out an output signal, X is a parameter representing the degree of non-linearity, and Y is a parameter which controls the inclination of the overall function. By adjusting the parameters X and Y, the frequency response of the band-limited signal can be changed.

In this way, the processed image signal S _{proc is} created. The problem to be solved by this embodiment and the manner of solving the problem according to this embodiment will be explained below by way of example. When the filtering which is determined depending on the picture element density of the original image signal S _org is performed on the original image signal S _org using a filter, thereby obtaining unsharp image signals S _us k, and band-limited signals are obtained from the unsharp image signals S _us k as disclosed in Japanese Unexamined Patent Publication No. JP 10-63837, the band-limited signals may be in the same frequency bands regardless of the pixel density of the original image signal S _org . In this case, however, the frequency responses of the band-limited signals differ depending on the pixel density of the original image signal S _org .

7 shows the frequency response of the band-limited signals (S _org - S _us 1, S _us 1 - S _us 2, S _us 2 - S _us 3, S _us 3 - S _us 4, S _us 4 - S _us 5, S _us 5 - S _us 6) obtained by performing the filtering processing with the in 3A shown one-dimensional filter with respect to an original image signal with a pixel density of 10 lines / mm. 8th shows the frequency response of band-limited signals (S _org - S _us 1, S _us 1 - S _us 2, S _us 2 - S _us 3, S _us 3 - S _us 4, S _us 4 - S _us 5, S _us S - S _us 6) obtained by filtering with respect to an original image signal with a pixel density of 6.7 lines / mm in 3B durchdurchüührt is shown one-dimensional filter. 9 shows the frequency response of the band-limited signals (S _org - S _us 1, S _us 1 - S _us 2, S _us 2 - S _us 3, S _us 3 - S _us 4, S _us 4 - S _us 5, S _us 5 - S _us 6) obtained by subjecting a filter processing to the in. _FIG. 6) with respect to an original image signal having a pixel density of 5 lines / mm 3A shown one-dimensional filter is performed. Like from the 7 to 9 is apparent, although the maximum frequency band of the band-limited signals is determinable according to the pixel density, however, the peak frequencies of the band-limited signals obtained from the respective original image signals except the maximum frequency band are reduced by 1/2 to 1/2, for example 1 ( only in the case of a pixel density of 10 lines / mm); 0.5; 0.25; 0.125 ... That is, the peak frequency of the k-th band-limited signal is 1/2 of the (k-1) th band-limited signal for each pixel density. The response of each peak frequency, that is the energy of each band-limited signal, differs depending on the pixel density. Thus, when the same parameters X and Y (formula (6)) are used for different pixel densities, the frequency response of the obtained processed image signal S _proc differs depending on the pixel density of the original image signal S _org .

According to the invention, to overcome this problem, the transformation function defining means 4 the parameters X and Y of the transformation function based on the pixel density information M supplied through the pixel density information input device 5 is input, and the band-limited signals are transformed in accordance with the transformation function defined depending on the pixel density of the original image signal S _org . 10 shows in detail the structure of the transformation function defining means 4 , As in 10 is shown contains the transformation function defining means 4 a parameter setting device 7 and a memory 8th in which combinations of the parameters X and Y are stored for each pixel density.

An example of the combination of parameters X and Y is shown in Table 1 below. In the combination of the parameters X and Y according to Table 1, the parameters X and Y for the pixel densities of 6.7 lines / mm and 5 lines / mm are set so that a processed image signal S _proc with a frequency response substantially equal to that which is obtained from the parameters X and Y for the pixel density of 10 lines / mm. In Table 1, the parameters X and Y apply to each pixel density in the upper lines for band-related signals in higher frequency bands, and since there is no band-limited signal with a frequency peak of 0.03 cycles / mm for the pixel densities of 6.7 lines / mm and 5 lines / mm, no parameter is specified.

Table 1

The parameter setting device 7 selects the combinations of parameters X and Y that are in memory 8th are stored, depending on the pixel density represented by the pixel density information M which is received from the pixel density information input device 5 is entered, and the selected combinations of parameters X and Y are entered into the transformation function defining means 4 entered. This defines the transformation functions based on the selected combinations of the parameters X and Y, and the non-linear processing means 3 performs the calculation according to formula (2) with respect to the band-limited signals on the basis of the transformation functions thus defined.

11 _{FIG. 12} shows the frequency response of the processed image signal S _proc obtained by processing the band-limited signals obtained from the 10-line / mm original image signal using the transform functions defined in the above-described manner. 12 _{FIG. 12} shows the frequency response of the processed image signal S _proc obtained by processing the band-limited image signals obtained from the original image signal of 6.7 lines / mm using the transform functions defined above. 13 _{FIG. 12} shows the frequency response of the processed image signal S _proc obtained by processing the band-limited signals obtained from a 5 line / mm original image signal using the transform functions defined in the above-described manner. Like from the 11 to 13 is removed, when the frequency emphasis is performed using the transform functions defined by the parameters X and Y in Table 1, the frequency response of the obtained processed image signal S _proc is substantially constant regardless of the pixel density of the original image signal.

The Parameters X and Y can For example, be set according to the method, which in Japanese unaudited Patent publication No. JP 10-63838. This procedure is explained below explained become. As a system for setting the parameters X and Y, there are a system that prompts the user for a desired one Frequency response over to enter a screen to see which one from the input Frequency response the user wants. In particular contains the system software, an input system, a display and the like.

When the parameters X and Y are set, a screen will show 7a ( 14 ) the frequency response of the processed image signal S _proc , which is obtained when the band-limited signals are processed by the current transform functions (the initial transform functions). The curve shows six moving points. That is, when the image processing is non-linear processing using n band-limited signals, the number of movable dots is n.

The user moves the required point or points to a location corresponding to a desired frequency response by operating a pointing device such as a mouse. The screen 7b in 14 shows this process. It runs so that the answers in the respective frequency bands are equal to each other. After moving the required points, the new points are recognized as identifying the desired frequency responses at the frequencies indicated by the points, and the frequency response curve is recalculated to indicate a new characteristic passing through the new points. The in 14 screen shown 7c illustrates this was standing. If a new characteristic corresponds to the desired characteristic, the user selects the in 7d shown "OK" button or otherwise selects the "Delete" button to carry out the parameter setting again. The parameters can be set in different ways. For example, the parameter setting means may be configured such that a plurality of frequency values are displayed on the screen in a sequence and the user inputs a desired frequency response for the displayed frequency.

If the instruction is to be made for a parameter setting, the input desired frequency response characteristics a1 to a6 (in the case where the image processing is to be performed using six band-limited signals) are set in the following simultaneous equations (7), for example: a1 = Y1 × S11 + Y2 × S21 + y3 × S31 + Y4 × S41 + Y5 × S51 + Y6 × S61 a2 = Y1 × S12 + Y2 × S22 + Y3 × S32 + Y4 × S42 + Y5 × S52 + Y6 × S62 a3 = Y1 × S13 + Y2 × S23 + y3 × S33 + Y4 × S43 + y5 × S53 + Y6 × S63 a4 = Y1 × S14 + Y2 × S24 + Y3 × S34 + y4 × S44 + Y5 × S54 + Y6 × S64 a5 = Y1 × S15 + Y2 × S25 + Y3 × S35 + Y4 × S45 + y5 × S55 + Y6 × S65 a6 = Y1 × S16 + Y2 × S26 + Y3 × S36 + Y4 × S46 + y5 × S56 + y6 × S66 Y1 ~ Y6 ≥ 0 (7) where a1 to a6 represent the desired frequency response, S11 to S66 represent the frequency response of the band-limited signals, created on the basis of the original image signal, and Y1 to Y6 are parameters indicating the slopes of the transformation functions near 0. The degree of the equations is determined by the number of band-limited signals, the above equations being given by way of example only.

By solving the simultaneous equations (7) with the target frequency response a1 to a5 and the frequency response of the band-limited signals S11 to S66 as known values as well as the parameters Y1 to Y6 as variables, the parameters Y1 to Y6 are obtained, and the transformation functions become the basis of the parameters defined. The parameters X representing the degree of non-linearity can be set to arbitrary values corresponding to the parameters Y1 to Y6, and can be set to, for example, fixed values corresponding to the pixel densities in Table 1 above. The parameters X and Y depending on the pixel density are in memory 8th saved.

The Values for the frequency response of the band-limited signals S11 to S66 are valid for six band-limited signals, where S11 to S16 (S21 to S26, ... S61 to S66) the frequency response for represent six frequencies of a bandlimited signal. The values of the frequency response of the band-limited signals S11 to S66 can be won by a transformation process, which usually for the frequency analysis based on the filter coefficients and the interpolation coefficient is used for the generation of the blur image signals.

The Parameters X and Y can be set by trial and error.

The following is based on the in 15 shown flow diagram, the operation of the image processing system 1 explained this embodiment. First, an original image _signal S _{org is input} to the image processing system 1 input, for example from an image reader (step S1). The original image _signal S _org is _input to the blur image signal generator 2 are input, and according to the picture element density of the original picture signal S _org , the blur image signals S _us k are generated (step S2). The pixel density information M with respect to the original image signal S _org is input to the pixel density information input device 5 entered (step S3). Then in the transformation function defining means 4 select combinations of the parameters on the basis of the pixel density information M in the above-described manner, and the transformation functions are defined on the basis of the selected parameter combinations (step S4). The steps S3 and S4 may be executed before the steps S1 and S2.

If steps S3 and S4 on the one hand and steps S1 and S2 on the other hand are carried out in parallel, the working time can be shortened. On the basis of the blur image signals S _us k, frequency-limited signals are generated, and are generated by the non-linear processing means 3 on the basis of the transform function defining means 4 defined transformation functions the nichtli neare processing (frequency emphasis) according to the above formula (2), whereby a processed image signal S _{proc is} obtained (step S5).

The frequency emphasis processing serves to form addition signals to be added to the original image signal to emphasize desired frequency components, for which the blur image signals are used. The frequency bands are differentially processed so that the signals in the respective frequency bands forming the individual addition signals become the desired signals to inhibit the generation of an artifact due to the frequency emphasis processing. In order to generate the desired signals, it is preferred if sigals have the desired frequency response independent of the _pixel density of the original image signal S _org . In the image processing system 1 In this embodiment, the parameters of the transformation functions are defined such that the frequency response of the processed image signal S _proc becomes substantially constant regardless of the pixel density of the original image signal S _org , and thus a processed image signal S _{proc of} substantially constant frequency response independent of the pixel density of the original image signal S _org can be won.

Although, in the above-described embodiment, the frequency emphasis processing represented by the above formula (2) is performed as non-linear processing, similarly, as non-linear processing, dynamic range compression and frequency emphasis processing may be performed simultaneously. 16 FIG. 11 is a view of a system for performing frequency emphasis and dynamic range compression processing as non-linear processing in an image processing system according to a second embodiment of the invention. FIG. As in 16 are shown, differences between two signals from the original image signal S _org and the blur image signals S _us k, by the filter means 10 and the interpolator 11 were formed, generated, whereby a plurality of band-limited signals are formed, which represent the frequency components in limited frequency bands of the original image signal (e.g., S _org - S _us 2 and the like - S _us 1, S _us 1). The in the filter device 19 The filters used are determined in accordance with the pixel density of the original image signal in the manner similar to the manner described above. The thus obtained band-limited signals are in the first and the second transformation part 3a and 3b entered and from respective transformation facilities 22a and 22b in the first and the second transformation arrow 3a and 3b processed.

The of the transformation device 22a in the first transformation part 3a The transformation performed is based on the use of the transform functions defined on the basis of the pixel density information M with respect to the original image signal S _org , as explained above. For example, the transformation functions according to 17 or 18 or combinations of those in the 17 and 18 used transformation functions. These transformation functions are obtained by adjusting the parameters such that the frequency response of the processed image signal S _proc becomes substantially constant regardless of the pixel density of the original image signal S _org .

According to the in 17 As shown, when the frequency-band of the band-limited signal is higher, the band-limited signals are transformed so as to suppress those of large amplitude and to increase the amount of suppression. This is because the fact is to be taken into account that the higher-frequency components in the edges of a current radiation image have a lower amplitude than the low-frequency components. In current radiation images, even a sharp edge does not have the shape of a correct step, the amplitude often decreases as the frequency increases. Thus, it is preferable that the suppression be from a smaller amplitude as the frequency of the band-limited signals becomes larger. In the 17 Functions shown serve this purpose.

In the 18 The transformation functions shown are for transforming the band-limited signals into those having values not greater than the absolute values of the band-limited signals, which values are determined based on the absolute values of the band-limited signals. As the frequency of the frequency band to be processed by the function becomes lower, the absolute value of the transformed picture signal obtained by transforming a band-limited signal having an absolute value in a predetermined range near 0 becomes smaller. That is, all functions pass through the origin and all have slopes of less than 1. Also, the slope near 0 is smaller as the frequency of the frequency band to be processed by the function becomes lower. When an addition signal obtained by adding up the transformed image signals is added to the original image signal S _org , these functions contribute to smoothing the connection point between the original image signal S _org and the addition signal, that is, smoothing the rise of the signal.

Similarly, the transformation will be from each transformation facility 22b in the second transformation part 3b using the in 19 or 18 shown transformation functions or kombinert executed from these transformation functions. The transformed image signals formed by the first and second trans formation parts 3a and 3b are issued to the operators 23a and 23b entered. The operator 23a performs the process of generating signals necessary for the frequency emphasis, and the operator 23b performs the operation of generating signals required for the dynamic range compression processing.

The operator 23a performs a frequency accentuation similar to that discussed above in connection with 6 was explained. That is: that of the first transformation part 3a transformed band-limited signals are added up, and an addition signal thus obtained is multiplied by an emphasis coefficient β determined according to the value of the original image signal S _org .

The operator 23b performs dynamic range compression processing in the following manner: that of the transform part 3b transformed, band-limited signals are added, and the addition signal thus obtained is subtracted from the original image signal S _org . Then, the difference signal thus obtained is transformed on the basis of a transform function to obtain a dynamic range compression coefficient. The by the operators 23a and 23b obtained signals are from an adder 28 is added to the original image signal S _org , whereby a processed image signal S _{proc is} obtained.

The above-described processing is described by the following formula (3): S proc = S org + β (S. org ) × F usm (S org , P us1 , P us 2, ..... S us N) + D {p org - F drc (S org , P us 1, p us 2, ..... S us N)} F usm (S org , P us 1, p us 2, ..... S us N) = {f u1 (S org - p us 1) + f u2 (S us 1 - p us 2) + ... + f uk (S us k - 1 - p us k) + ... + f U.N (S us N - 1 - S us N) F drc (S org , P us 1, p us 2, ..... S us N) = {f d1 (S org - p us 1) + f d2 (S us 1 - p us 2) + ... + f dk (S us k - 1 - p us k) + ... + f dN (S us N - 1 - S us N) (3) wherein S _proc is a processed image signal, S _org, an original image signal, S _us k (k = 1 to N) is an unsharp image signal, f _uk (k = 1 to N) is a transformation function used in the first transformation section, f _dk (k = 1 to N) is a transformation function used in the second transformation part, β (S _org ) is an emphasis _coefficient determined on the basis of the original _{image signal} , and D {S _org -F _drc (S _org , S _us 1, S _us 2, ... S _us N)} is a dynamic range compression coefficient obtained based on the low-frequency component _signal , and D is a function of transforming D {S _org -F _drc (S _org , S _us 1, S _us 2 , ... S _us N)} is.

In the in 16 In the processing system shown, the addition signal used in the frequency emphasis and the signal used in the dynamic range compression processing are separately obtained on the basis of the original image signal S _org and finally added. However, it is possible to first perform the frequency emphasis or the dynamic range compression processing on the original image signal S _org and then perform the other processing on the resultant signal. In general, the low density part of a radiation image contains a relatively large amount of noise because the low density part is exposed to less radiation when taking the radiation image. Since the frequency emphasis processing is a density-dependent processing in which the amount of emphasis increases with increasing density, the low-density portion is not amplified, and hence the noise is not amplified even when the frequency emphasis is made directly with respect to the original image signal. On the other hand, when the frequency emphasis is performed after the dynamic range compression processing has been performed on the original image signal, the low-density part is further increased in density by the dynamic range compression processing, and hence the low-density part is enhanced by the subsequent frequency emphasis, which also Noise is amplified. Consequently, it is preferable that each signal is obtained on the basis of the original image signal S _org . In view of saving processing time, it is also preferable if the two processes are performed in parallel.

By defining the parameters for the transform functions such that the frequency response of the processed image signal S _proc becomes constant independently of the pixel density of the original image signal S _org , the processed image signal S _{proc can become} a substantially constant frequency response from the pixel density of the original image signal S _org .

In the second embodiment as described above, the frequency response of the processed image signal S _{proc is} as shown in _FIG 20 is shown when the parameters of the transformation functions are set as shown in the following table 2 in the case where the pixel density of the original image signal is 10 lines / mm. As in the 7 to 9 In the case of an original image signal of 10 lines / mm, a band-limited signal corresponding to 1 cycle / mm exists, but in the case of an original image signal of 6.7 lines / mm or 5 lines / mm, there is no band-limited signal which is 1 cycle / mm corresponds. Thus, in the case of original picture signals of 6.7 lines / mm and 5 lines / mm, the frequency response of the processed picture signals S _proc can not coincide with the processed picture signal S _proc for the original picture signal of 10 lines / mm in the entire frequency bands. In an image obtained by reproducing a processed image signal, the low-frequency components are visually more clearly discernible than the high-frequency components. Thus, in this case, it is preferable that the transform functions for band-limited signals of 6.7 lines / mm and 5 lines / mm are defined by the parameters X and Y shown in Table 2 so that the high-frequency components are not less than 1 cycle / mm coincide with those of the band-limited signals at 10 lines / mm to a certain extent, and the low-frequency components below 1 cycle / mm coincide with those of the band-limited signals at 10 lines / mm. When the parameters of the transform functions shown in Table 2 below are set in the case where the pixel densities of the original image signals are 6.7 lines / mm and 5 lines / mm, the frequency response of the processed image signals S _proc becomes as shown in _FIGS 21 and 22 is shown. As can be seen there, when 6,7 line / mm and 5 line / mm original picture signals are processed using the transformation functions defined by the parameters shown in Table 2, the frequency response of the processed picture signals S _{proc is in} good agreement with the processed picture signal S _proc for the original image signal of 10 lines / mm, which concerns the low-frequency components below 1 cycle / mm, although little match with the processed image signal S _proc for the original image signal of 10 lines / mm in the higher-frequency components of not less than 1 cycle / mm is present. Consequently, the frequency response which is constant regardless of the pixel density of the original image signal at least at the low-frequency components which are more clearly visible than the high-frequency components can be obtained.

Table 2

Although in the above-explained embodiments, the original image signal S _{org is} an image signal read from a stimulable phosphor sheet, the original image signal S _{org may be} any image signal, provided the resolution of the image represented by the image signal, that is, the relation between object size and sampling intervals (dpi), is known.

Although in the embodiments described above non-linear functions are used and a non-linear Processing regarding the band-limited signals is performed, linear functions or Constant can be used as transformation functions. In one In such case, the parameters of the linear functions or the Values of the constants corresponding to the pixel density of the original image signal set.

Although in the embodiments described above, the blur image signals are from the original image signal obtained by filtering and interpolation / enlargement, and the band-limited signals are obtained from the original image signal and the blur image signals, the band-limited signals may be generated by, for example, transforming the original image signal into multi-resolution image signals using a wavelet transform or a Laplace pyramid; wherein the blur image signals are generated from the image signals of the respective resolutions, and the band-limited signals are again obtained from the blur image signals.

Although in the embodiments described above, in 10 illustrated parameter setting device 7 is configured such that the parameters X and Y are set on the basis of the original image signal and its pixel density each time an original image signal S _{org is} input, the parameter setting means 7 also be designed so that the parameters X and Y are set for the original image signal S _org by comparing the picture element density of the original image signal S _org with reference picture element densities stored in the memory 8th are stored together with the parameters X and Y for the individual pixel densities, and as parameters X and Y for the original image signal S _org, the parameters X and Y are set for the reference _{pixel density} closest to the pixel density of the original image signal S _org . The closest to the pixel density of the original image signal S _org reference pixel density can be determined, for example, in the following manner: assuming that in the memory 8th Let a, b and c be (a <b <c) together with the parameters X and Y for the respective reference pixel densities, and let the pixel density of the original image signal S _org be x, and the reference pixel density be c if x> Let b + (c - b) / 2 be b if b + (c - b) / 2 ≥ x> a + (b - a) / 2 and let a be if a + (b - a) / 2 ≥ x, there can be at least two reference pixel density for 10 lines / mm, 6.7 lines / mm and 5 lines / mm. In this case, if the pixel density of the original image _signal S _proc to be processed is larger than 6.7 + (10 - 6.7) / 2 = 8.35, the combinations of the X and Y parameters for the pixel density become 10 lines / mm in accordance with Table 1 or Table 2 as the parameters X and Y for the original image _signal S _org . When the pixel density of the original image signal S _org to be processed is not larger than 6.7 + (10 - 6.7) / 2 = 8.35 and higher than 5 + (6.7 - 5) / 2 = 5.85, thus, the combinations of parameters X and Y for the pixel density of 6.7 lines / mm are set as the parameters X and Y for the original image _signal S _org according to Table 1 or Table 2. When the pixel density of the original image signal S _org to be processed is not larger than 5 + (6.7 - 5) / 2 = 5.85, the combinations of the parameters X and Y for the pixel density of 5 lines / mm become as shown in Table 1 or 2 is set as the parameters X and Y for the original image _signal S _org .

It is preferred if the parameters selected for storage relate to the processed image signal S _proc and / or the pixel density of the original image signal S _org . In this embodiment, when an image signal is output once more (as a hard copy is output on a photographic film or displayed on a screen), an image having a quality corresponding to that of the previously output image can be obtained. The selected parameters and the pixel density of the original image signal S _org may be stored in the same storage medium or in different storage media.

An image processing system according to a third embodiment of the invention will now be described. An image processing system 101 According to the third embodiment of the present invention, it is also suitable for performing frequency emphasis using blur image signals with respect to the original image signal obtained by reading a radiation image of a human body recorded on a stimulable phosphor sheet so that an image capable of diagnosis is obtained can. An image reproduced on the basis of the processed image signal is recorded primarily on a photographic film and used for diagnosis.

As in 23 can be seen is the image processing system 101 according to the third embodiment of the invention substantially the same as the image processing system 1 according to the first embodiment of the invention. Consequently, wear in 23 such elements according to those 1 are equivalent, like reference numerals, and will not be described in detail here. The image processing system of this embodiment includes a blur image signal generator 2 , a processing device 3 which performs frequency emphasis for emphasizing a specific frequency and provides a processed image signal S _proc , a transformation function calculator 104 and a resolution information input device 105 , The image processing system of the third embodiment is different from that of the first embodiment mainly by the configuration of the transformation function calculating means. That is, in this embodiment, the transformation function calculating means calculates 104 those transformation functions which the processing device 3 used in frequency emphasis by correcting reference transformation functions, which are used to emphasize the frequency with respect to an original picture signal having a known reference picture element density depending on the resolution of the original picture signal S _org and represented by the resolution information M _transmitted through the resolution information input device 105 is entered, which is clear from the following explanation. The transformation functions for the original image signal were previously determined and stored.

Blur image signals S _usk are determined in the same manner as described above in connection with 2 has been explained, and the frequency emphasis is performed using the blur image signals S _usk in the same manner as described above in connection with FIG 6 was explained.

In this embodiment, the transformation functions f _k (k stands for 1 to N) for transforming the band-limited signals are determined in the following manner: in this particular embodiment, the transformation functions f _{k are} constants, the reference resolution 10 Lines / mm. 24 shows the frequency response of the band-limited signals obtained from the original image signal S _org with a resolution of 10 lines / mm. As in 24 can be seen, six band-limited signals are obtained from the original image signal S _org . The original image signal S _org with a resolution of 10 lines / mm has a Nyquist frequency of 5 cycles / mm. The peak frequency of the band-limited signal in the highest frequency band coincides with the Nyquist frequency. The peak frequency of the band limited signal in the second highest frequency band is 1.0 cycle / mm, namely 1/5 of the Nyquist frequency. The peak frequency of the band-limited signal is reduced by 1/2, for example, to 0.5 cycles / mm, 0.25 cycles / mm, 0.12 cycles / mm, and 0.06 cycles / mm, when the frequency band of the band-limited signal becomes smaller becomes. In this particular embodiment, the transform functions f _k (k stands for 1 to N) for transforming the band-limited signals obtained from the original image signal S _{org to} a resolution of 10 lines / mm are as shown in the following Table 3:

Table 3

In order to accentuate or emphasize a desired frequency component of the original image signal S _org , it is necessary to perform the frequency emphasis by using transformation functions f _k which are fixed frequency by frequency of the band-limited signals. However, if the transformation functions are used for an original image signal S _org with a resolution of 10 lines / mm as they are, the frequency response of the processed image signal S _proc differs for a frequency emphasis of an original image signal S _org with a resolution of 300 dpi from that of the processed image signal S _proc obtained from the original image signal S _org at a resolution of 10 lines / mm. In 25 the solid line shows the frequency response of the processed image signal S _proc obtained by performing the frequency emphasis using the transformation functions shown in Table 3 with respect to the original image signal S _org at 10 lines / mm, the broken line shows the frequency response of the processed image signal S _proc , obtained by performing frequency emphasis using the transformation functions shown in Table 3 with respect to the original image signal S _org at 300 dpi.

To overcome this problem, when the original image S _{org is} inputted at a resolution other than 10 lines / mm (reference resolution), the transformation functions shown in Table 3 (the reference transformation functions) corresponding to the resolution (300 dpi in this particular embodiment) Original _{picture signal} S _org corrected.

26 shows the frequency response of the band-limited signals obtained from an original image signal S _org with a resolution of 300 dpi. How out 26 is apparent, the original image signal S _org at 300 dpi is 5.9 cycles / mm in the Nyquist frequency, which coincides with the peak frequency of the band-limited signal in the highest frequency band. The peak frequency of the band-limited signal in the second highest frequency band is 1.018 cycles / mm, ie 1/5 of the Nyquist frequency. The peak frequency of the band-limited signal decreases in 1/2 increments, for example, 0.59 cycles / mm, 0.30 cycles / mm, 0.15 cycles / mm, and 0.07 cycles / mm when the frequency band of the band-limited signal gets lower. This relation is substantially coincident with the relation for the original image _signal S _org at 10 lines / mm according to FIG 24 , and independently applies to the resolution of the original image signal S _org as long as the filters used in the generation of the low-frequency image signals are not changed.

The transformation functions for transforming the band-limited signals obtained from a 300 dpi original image signal S _org are obtained by taking the reference transformation functions shown in Table 3 on the basis of the peak frequencies of the 10 lines / mm original image signal S _org (reference original image signal S _org 0) and those of the band-limited signals obtained from the original image signal S _org at 300 dpi are corrected. This correction can be done, for example, by interpolation or extrapolation. For example, the transform function for the band-limited signals whose peak frequency is 1.18 cycles / mm is obtained by determining the relation between the peak frequencies and the values of the transformation functions based on the values of the transformation functions at 5 cycles / mm and 1 cycle / mm in Table 3, wherein the value of the peak frequency (1.18) is substituted into the following formula (6) representing the predetermined relation between the peak frequencies and the values of the transformation functions: fx = 1/40 x C + 0.875 (6), where fx = the value of the transformation function and C represents the peak frequency. Such a linear interpolation is performed on all the peak frequencies of the band-limited signals obtained from the original picture signal S _{org of} 300 dpi, to thereby obtain the transform functions for all the band-limited signals, as shown in Table 4 below:

Table 4

Although according to the above Description of a linear interpolation use, can also other interpolations are used.

In 27 The solid line shows the frequency response of the processed image signal S _proc obtained by emphasizing the frequency using the transformation functions shown in Table 3 with respect to the original image signal S _{org of} 10 lines / mm, and the broken line shows the frequency characteristic of the processed image signal S _proc was obtained by the frequency _{emphasis processing} using the transform functions for the original image signal S _org at 300 dpi shown in Table 4. How out 27 It can be seen that the processed image signals S _{proc are} approximately equal in frequency response.

The following is based on the in 28 illustrated flow chart, the operation of the image processing system 101 explained this embodiment. First, an original image _signal S _{org is input} to the image processing system 101 input, for example from an image reading device (step S11). The original image _signal S _org is _input to the blur image _signal generator 2 are inputted, and blur image signals S _us k are generated therefrom (step S12). Resolution information M relating to the original image _signal S _org is input to the resolution information input device 105 entered (step S13). Then, the transformation function calculating means calculates 4 the transformation functions based on the resolution information M in the same manner as explained above (step S14). The steps S3 and S4 may be executed before the steps S1 and S2. If steps S3 and S4 and steps S1 and S2 are executed in parallel, the processing time can be shortened. Frequency-limited signals are generated based on the blur image signals S _us k, and the frequency emphasis represented by the above formula (2) is provided by the processing means 3 based on the transformation functions provided by the transformation function calculator 4 were calculated to obtain a processed image signal S _proc (step S15).

The frequency emphasis processing (frequency emphasis) is for generating addition signals which are added to the original image signal to emphasize desired frequency components by making use of the blur image signals. The frequency bands are processed in different ways so that the signals in the respective frequency bands forming the individual addition signals become the desired signals to inhibit the generation of an artifact due to the frequency emphasis processing. In order to generate the desired or desired signals, it is preferred that signals having a desired frequency response be used independently of the resolution of the original image signal S _org . In the image processing system 101 This embodiment uses the transformation functions is calculated so that the frequency response of the processed image signal S _proc becomes substantially constant regardless of the resolution of the original image signal S _org , and accordingly, a processed image signal S _proc having a substantially constant frequency response is obtained irrespective of the resolution of the original image signal S _org .

In addition, since the original image signal transformation functions S _org having resolutions other than the reference resolution are obtained by correcting the reference transformation functions, it is not necessary to provide a plurality of different resolution transformation functions, thereby avoiding troubles in handling a variety of transformation functions ,

Although in the third embodiment described above, only a series of reference transformation functions for a reference resolution of 10 lines / mm is prepared and transformation functions for other resolutions are calculated on the basis of these one series of reference transformation functions, it is possible to make a plurality a series of reference transform functions for a plurality of resolutions, which are relatively commonly used to compute object pixel density transform functions based on one of the reference resolutions. For example, when first and second reference transformation functions are respectively prepared in advance for first and second resolutions, transformation functions for an object-to-be-processed image signal based on the first reference transformation functions can be calculated when the resolution of the object original image signal is less than the middle between the first and second resolutions, as indicated by the arrow A in FIG 29A on the other hand, based on the second reference transformation functions, when the resolution of the object original image signal is not less than the center between the first and second resolutions, as shown by the arrow B. Otherwise, the transformation functions for the object original image signal may be calculated on the basis of the first reference transformation functions, if the resolution of the object original image signal is not greater than the first resolution, as in FIG 29B is represented by the arrow C, and based on the second reference transformation functions, when the resolution of the object original image signal is greater than the first resolution, as indicated by the arrow D.

Although in the third embodiment described above, the transformation functions Constants are, so can they also be non-linear functions with predetermined slopes.

in the The following will be an image processing system according to a fourth embodiment described in the invention of non-linear functions Use is made.

In the fourth embodiment, as transformation functions f _k, non-linear functions are used, which are described by the following formula (8): f (S in ) = S out = S in × Y × {exp (X / S in ) - 1} / {exp (X / S in ) + 1} (8) wherein S is _in an input signal, S _out an output signal, X indicates a parameter indicating the degree of non-linearity, such as a suppression state, and Y indicates a parameter which controls the inclination of the overall function, that is the frequency response. By adjusting the parameters X and Y, the frequency response of the band-limited signal can be changed. In this embodiment, the parameters X and Y shown in Table 5 below are used. Table 5 shows combinations of parameters X and Y for resolutions of 10 lines / mm; 6.7 lines / mm or 5 lines / mm. In Table 5, the X and Y parameters apply to any higher-row resolution for band-related signals in higher frequency bands. As in 30 As shown, as the Nyquist frequency decreases, the number of band-applied signals used for nonlinear processing decreases. For an original picture signal of 10 lines / mm, for example, the number of band-limited signals is 6, while in the case of original picture signals of 6.7 lines / mm and 5 lines / mm, the number of band-limited signals is 5. This is because no band-stopped signal with a peak frequency exists at 0.03 cycles / mm for resolutions of 6.7 lines / mm and 5 lines / mm. When the reference resolution is 10 lines / mm, the transformation functions defined by combinations of the parameters X and Y in Table 5 are used as the reference transformation functions, and the resolution of the original image signal to be processed is 200 dpi, which will be described below.

Table 5

31 shows the frequency response of band-limited signals obtained from an original image signal Sog with a resolution of 200 dpi. How out 31 For the original image signal S _org at 200 dpi, the Nyquist frequency is about 4.0 cycles / mm (more precisely, 3.937 cycles / mm), which coincides with the peak frequency of the band-stopped signal in the highest frequency band. The peak frequency of the band-beat signal in the second highest frequency band is 0.8 cycles / mm, which is 1/5 of the Nyquist frequency. The peak frequency of the band-limited signal is reduced in steps of 1/2, for example 0.4 cycles / mm; 0.2 cycles / mm; 0.1 cycles / mm and 0.05 cycles / mm when the frequency band of the band-limited signal becomes lower.

This relation is substantially coincident with the relation for the original image _signal S _{org of} 10 lines / mm written in 24 is shown and is independent of the resolution of the original image signal S _org , as long as the filters used to generate the low resolution image signals are not changed. 32 _{FIG. 12} shows the frequency response of the processed image signal S _proc obtained by processing the original image signal S _org at 200 dpi using the reference transformation functions (dashed line) and the processed image signal S _proc 0 obtained by processing the reference original image signal S _org 0 with the reference resolution using the reference transformation functions (solid line). How out 32 is removed, the processed image signal S _proc in the frequency bands lower than that of the processed image signal S _proc is emphasized 0th

The transformation functions are thus obtained for transforming from an original image signal S _org at 200dpi obtained band-limited signals characterized in that the reference transformation functions of the original image signal S _org are corrected on the basis of the resolution as in the third embodiment. In this embodiment, the peak frequencies of the band-limited signals from the reference original image signal S _{org are} 10 lines / mm, the parameters of the reference transform function (the values Y in Table 5) and the peak frequencies of the band-limited signals from the original image signal S _{org are} 200 dpi as well as the number of transformation functions to be obtained (six). Thus, first, the responses at peak frequencies of the band-limited signals obtained from the original image signal S _org at 200 dpi are first obtained according to the formula (9), based on the frequency response of the band-limited obtained from the reference original image signal S _org at 10 lines / mm Signals, the parameter Y of the reference transformation functions and the peak frequencies of the band-limited signals obtained from the original image signal at 200 dpi. R1 = X1 [F1] Y1 + X2 [F1] Y2 + X3 [F1] Y3 + X4 [F1] Y4 + X5 [F1] Y5 + X6 [F1] Y6 R2 = X1 [F2] Y1 + X2 [F2] Y2 + X3 [F2] Y3 + X4 [F2] Y4 + X5 [F2] Y5 + X6 [F2] Y6 R3 = X1 [F3] Y1 + X2 [F3] Y2 + X3 [F3] Y3 + X4 [F3] Y4 + X5 [F3] Y5 + X6 [F3] Y6 R4 = X1 [F4] Y1 + X2 [F4] Y2 + X3 [F4] Y3 + X4 [F4] Y4 + X5 [F4] Y5 + X6 [F4] Y6 R5 = X1 [F5] Y1 + X2 [F5] Y2 + X3 [F5] Y3 + X4 [F5] Y4 + X5 [F5] Y5 + X6 [F5] Y6 R6 = X1 [F6] Y1 + X2 [F6] Y2 + X3 [F6] Y3 + X4 [F6] Y4 + X5 [F6] Y5 + X6 [F6] Y6 (9) where R1 to R6 represent responses at peak frequencies (200 dpi), X1 [F1] to X6 [F6] represent responses of the band-limited signals obtained from the reference original image signal at frequencies of F1 to F6, and Y1 to Y6 the parameters of Reference transformation functions are.

Subsequently, by solving the simultaneous equations 10 given below, the parameters A1 to A6 of the original image signal transformation functions of 200 dpi are obtained: R1 = Z1 [F1] * A2 + Z2 [F1] * A2 + Z3 [F1] * A3 + Z4 [F1] * A4 + Z5 [F1] * A5 + Z6 [F1] * A6 R2 = Z1 [F2] A1 + Z2 [F2] * A2 + Z3 [F2] * A3 + Z4 [F2] * A4 + Z5 [F2] * A5 + Z6 [F2] * A6 R3 = Z1 [F3] * A1 + Z2 [F3] A2 + Z3 [F3] * A3 + Z4 [F3] * A4 + Z5 [F3] * A5 + Z6 [F3] * A6 R4 = Z1 [F4] * A1 + Z2 [F4] * A2 + Z3 [F4] A3 + Z4 [F4] * A4 + Z5 [F4] * A5 + Z6 [F4] * A6 R5 = Z1 [F5] * A1 + Z2 [F5] * A2 + Z3 [F5] * A3 + Z4 [F5] A4 + Z5 [F5] A5 + Z6 F5 A6 R6 = Z1 F6 A1 + Z2 F6 A2 + Z3 F6 A3 + Z4 F6 A4 + Z5 F6 A5 + Z6 [F5] A6 (10) where R1 to R6 are the responses at peak frequencies (200 dpi) according to formula (9), Z1 [F1] to Z6 [F6] are responses of the band-limited signals obtained from the original image signal at 200 dpi at frequencies F1 to F6, and A1 to A6 are the parameters of the transformation functions for the original image signal at 200 dpi.

When the frequency emphasis is made with respect to the original image signal at 200 dpi by using the transform functions thus obtained, a processed image signal S _proc which is approximately constant in its frequency response can be obtained regardless of the resolution of the original image signal.

In the following, a method of calculating Z1 [F1] to Z6 [F6] in the formula (10) will be described. As stated above, the band-limited signals obtained from the reference original image signal S _org 0 at 10 lines / mm have the in 24 illustrated frequency response. The characteristic of the band-limited signal obtained in the highest frequency band, represented by the solid line, can be obtained by obtaining the frequency characteristic of the blur image signal by multiplying the Fourier-transformed value of the filter coefficient of in 3A shown one-dimensional filter with a Fourier-transformed value of the filter coefficient of in 5 The frequency response of the thus obtained blur image signal is subtracted from the frequency response of the reference original image signal S _org 0 (= 1 in the entire frequency bands). The characteristic of the band-limited signals in the lower frequency bands can be obtained by obtaining the frequency characteristic of the unsharp image signal based on the filter coefficient for obtaining the low resolution image signal corresponding to the frequency band and the filter coefficient for obtaining the unsharp image signal, whereupon the frequency response of the thus obtained unsharp image signal is subtracted from Frequency response of the blur image signal in the frequency band higher by one level.

When the response of the band-limited signal in the highest frequency response of the reference original image signal S _org 0 is represented by a function X1 [Fi] (Fi is a frequency) and values obtained by sampling the response of the band-limited signal in the highest frequency band of the reference original image signal 2048 frequencies is represented by a function X1 [I], the relation between Fi and i is according to the following formula (11): i = 2047 × Fi / fnq (11) where fnq is a Nyquist frequency and 2047 ≥ i ≥ 0. This is in 33 shown. If the function X1 [i] is represented as a table, i means an address. By obtaining responses for the band-limited signals in all frequency bands in the form of functions X1 [i] to X6 [i] and manipulating these functions as tables, values for X1 [F1] to X6 [F6] in the above formulas (9 ) and (10) win easily.

The responses Z1 [F1] to Z6 [F6] are obtained based on the functions X1 [i] to X6 [i]. The answer Z1 [Fi] at a given frequency Fi is obtained from the function X1 [i] as follows: The answer Z1 [Fi] is the response of the bandlimited signal in the highest frequency band within the bandlimited signals resulting from the original picture signal 200 dpi are obtained. The Nyquist frequency of the original picture signal of 200 dpi is 4 cycles / mm, and if the values obtained by sampling the response of the band-limited signal in the frequency response in which the answer Z1 [Fi] is 2048 frequencies are given by one function Z1 [i], as in the case of X1 [i], the relation between Fi and i is as represented by the above formula (11). 34 shows the relation between the functions X1 [i] and Z1 [i]. Although the functions X1 [i] and Z1 [i] in the concept are different functions different in the Nyquist frequency, the answer Z1 [Fi] can be approximated using the function X1 [i]. For example, the response Z1 [Fi] at 2 cycles / mm is obtained as follows: Since i = 2047 × 2/4 = 1024 according to the formula (11) above, the answer Z1 [Fi] is 2 cycles / mm as the value of the radio tion X1 [i] at the address 1024.

By Treat functions X1 [i] to X6 [i] as tables and through Determining the address i based on the formula (11) and the Nyquist frequency of the original picture signal, with which the frequency emphasis is to be carried out, and with reference to the tables (the functions X1 [i] to X6 [i]) corresponding to the address i, the answers Z1 [F1] to Z6 approximate [F6].

When the frequency emphasis is to be performed with respect to an original image signal S _org at 300 dpi, the responses at frequencies corresponding to the peak frequencies of the band-limited signals obtained from the original 300 dpi image signal are calculated according to the above formula (8) on the basis of the frequency response of the reference original image signal obtained band-limited signals, the peak frequencies obtained from the original image signal at 300 dpi band-limited signals and the parameter Y of the reference transformation functions won. Subsequently, the parameters for the transformation functions are obtained by the above formula (10). In the original picture signal of 300 dpi, the Nyquist frequency is 5.9 cycles / mm, which is higher than the Nyquist frequency of the reference original _picture signal S _org 0. Since X1 [F1] to X6 [F6] of the above formula (9) For the range of 0 to 5 cycles / mm, no value for 5.9 cycles / mm is given by the above formula (9), and the value R1 can not be determined. In this case, the peak frequencies of the band-limited signals obtained from the reference original picture signal are taken as the peak frequencies of the band-limited signals obtained from the original picture signal of 300 dpi.

In the following, the case will be explained in which the frequency _{emphasis is performed} with respect to an original image signal S _org at 100 dpi. Since the Nyquist frequency of the original picture signal of 100 dpi is about 2.0 cycles / mm (more specifically, 1.97 cycles / mm), the number of band-limited signals is 5. Accordingly, the above formulas (9) and (10) are shown in FIG The following formulas (12) and (13) are rewritten, and for the transformation functions are obtained according to the formulas (12) and (13) five parameters: R1 = X1 [F1] Y1 + X2 [F1] Y2 + X3 [F1] Y3 + X4 [F1] Y4 + X5 [F1] Y5 + X6 [F1] Y6 R2 = X1 [F2] Y1 + X2 [F2] Y2 + X3 [F2] Y3 + X4 [F2] Y4 + X5 [F2] Y5 + X6 [F2] Y6 R3 = X1 [F3] Y1 + X2 [F3] Y2 + X3 [F3] Y3 + X4 [F3] Y4 + X5 [F3] Y5 + X6 [F3] Y6 R4 = X1 [F4] Y1 + X2 [F4] Y2 + X3 [F4] Y3 + X4 [F4] Y4 + X5 [F4] Y5 + X6 [F4] Y6 R5 = X1 [F5] Y1 + X2 [F5] Y2 + X3 [F5] Y3 + X4 [F5] Y4 + X5 [F5] Y5 + X6 [F5] Y6 (12) where R1 to R5 represent responses at peak frequencies (100 dpi), X1 [F1] to X6 [F5] are responses of the band-applied signals obtained from the reference original image signal at frequencies F1 to F5, and Y1 to Y6 the parameters of the reference Are transformation functions. R1 = Z1 [F1] * A1 + Z2 [F1] * A2 + Z3 [F1] * A3 + Z4 [F1] * A4 + Z5 [F1] * A5 R2 = Z1 [F2] * A1 + Z2 [F2] A2 + Z3 [F2] * A3 + Z4 [F2] * A4 + Z5 [F2] * A5 R3 = Z1 [F3] * A1 + Z2 [F3] * A2 + Z3 [F3] * A3 + Z4 [F3] A4 + Z5 [F3] A5 R4 = Z1 F4 A1 + Z2 F4 A2 + Z3 F4 A3 + Z4 F4 A4 + Z5 F4 A5 R5 Z1 F5 · A1 + Z2 [F5] · A2 + Z3 [F5] · A3 + Z4 [F5] · A4 + Z5 [F5] A5 (13) where R1 to R5 are responses at peak frequencies (100 dpi) obtained in accordance with the formula (12), Z1 [F1] to Z5 [F5] are responses of the band-fed signals obtained from the original image signal at 100 dpi at frequencies F1 to F5 and A1 to A5 are the parameters of the transformation functions for the original image signal at 100 dpi.

Although the frequency emphasis according to the formula (2) is performed as transformation processing in the third and fourth embodiments described above, simultaneously with the transformation processing, dynamic range compression processing and frequency emphasis may be performed. 35 FIG. 10 is a view of a system for performing frequency emphasis processing and dynamic range compression processing as transformation processing in an image processing system according to a fifth embodiment of the invention. FIG. As in 35 2, differences between two signals, namely the original image signal S _org and a blur image signal S _us k, are detected by the filter means 100 and the interpolator 111 are obtained, which represent frequency components in limited frequency bands of the original image signal is obtained, whereby a plurality of band-limited signals (S _us 2 and the like, for example, S _org - - S _us 1, S _us 1). The band-limited signals thus obtained are converted into a first and a second transformation part 103a and 103b entered and by relevant transformation facilities 122a and 122b within the first or the second transformation part 103a and 103b processed.

The of the transformation device 122a in the first transformation part 103a The performed transformation is performed by using transformation functions calculated on the basis of the resolution of the original image _signal S _org in the manner described above. For example, as reference transformation functions, the in 36 or 37 transformation functions based on the reference transformation functions are calculated depending on the resolution of the original image signal S _org to be processed such that the frequency response of the processed image signal S _{proc is} substantially constant regardless of the resolution of the original image signal S _org becomes.

According to the in 36 As shown, when the frequency band of the band-limited signal is higher, the band-limited signals are transformed so as to suppress those having a larger amplitude and to increase the amount of suppression. This is to take account of the fact that higher-frequency components contained in the edges of the current radiation image have a smaller amplitude than low-frequency components. With current radiation images, at higher frequencies, the amplitude often becomes smaller even if there is no sharp edge in the form of a correct step. Thus, it is preferable that the suppression be from a smaller amplitude as the frequency of the band-limited signals becomes larger. In the 36 Functions shown serve this purpose.

In the 37 The transformation functions shown are used to transform the band-limited signals into signals whose values are not greater than the absolute value of the band-limited signals, the values being determined on the basis of the absolute values of the band-limited signals. As the frequency of the frequency band to be processed by the function becomes smaller, the absolute value of a transformed picture signal obtained by transforming a band-limited signal having an absolute value in a predetermined range near 0 becomes smaller. That is, the functions all run through the origin and all have slopes of less than 1. Also, the slope near 0 is smaller as the frequency of the frequency band to be processed by the function decreases. When an addition signal obtained by adding up the transformed image signals is added to the original image signal S _org , these functions contribute to smoothing the transition between the original image signal S _org and the addition signal, that is, an increase of the signal.

Similarly, the transformation is done by the transformation engine 22b in the second transformation part 103b by using the in 38 or 37 shown transformation functions or combinations thereof. That of the first and the second transformation part 103a and 103b output, transformed image signals become relevant operators 123a respectively. 123b entered. The operator 123a performs an operation that generates signals required for the frequency emphasis processing and the operator 123b performs an operation to generate signals required for the dynamic range compression processing.

The operator 123a performs a similar frequency emphasis as in the third embodiment described above. That is: that of the first transformation part 103a transformed band-limited signals are added up, and an addition signal thus obtained is multiplied by an emphasis coefficient β determined according to the value of the original image signal S _org . The operator 123b performs dynamic range compression processing in the following manner: that of the second transformation part 103b transformed band-limited image signals are added, and the addition signal thus obtained is subtracted from the original image signal S _org . Then, the thus obtained difference signal is transformed based on a transform function to obtain a dynamic range compression coefficient.

The by the operators 123a and 123b The received signals are added to the original image signal S _org by an adder 128 to obtain a processed image signal S _proc .

The above-described processing is described by the following formula (3): S proc = S org + β (S. org ) × F usm (S org , P us 1, p us 2, ... S us N) + D {p org - F drc (S org , P us 1, p us 2, ... S us N)} F usm (S org , P us 1, p us 2, ... S us N) = {f ul (S org - p us 1) + f u2 (S us 1 - p u2 2) + ... + f uk (S us k - 1 - p us k) + ... + f (S us N - 1 -S us N) F drc (S org , S us 1, p us 2, ..... S us N) = {f d1 (S org - p us 1) + f d2 (S us 1 - p us 2) + ... + f dk (S us k - 1 - p us k) + ... + f dN (S us N - 1 - S us N) (3) wherein S _proc is a processed image signal, S _org, an original image signal, S _us k (k = 1 to N) is an unsharp image signal, f _uk (k = 1 to N) is a transformation used in the first transformation section, f _dk (k = 1 to N) is a transformation function used in the second transformation part, β (S _org ) is an emphasis _coefficient determined on the basis of the original _{image signal} , and D {S _org -F _drc (S _org , S _us 1, S _us 2, .. .. S _us N)} is a dynamic range compression coefficient that is determined based on the low-frequency component _signal and D is a function for transforming D {S _org -F _drc (S _org , S _us 1, S _us 2, .... S _us N)} is.

At the in 35 In the image processing system, the signal used in the frequency emphasis and the signal used in the dynamic range compression processing are separately obtained on the basis of the original image signal S _org and finally added. However, it is also possible to first perform either the frequency emphasis or the dynamic range compression processing on the original image signal S _org , and then subject the resulting signal to further processing. Basically, the lower density portion of a radiation image contains a relatively large amount of noise because the lower density range is exposed to less radiation when taking the radiation image. Since the frequency emphasis is a density-dependent processing in which the amount of emphasis increases with increasing density, the low-density region is not emphasized, and hence the noise is not emphasized even when the frequency emphasis is directly related to the original image signal. On the other hand, when the frequency emphasis is performed after the dynamic range compression processing has been performed on the original image signal, the density-reduced portion is increased in density by the dynamic range compression processing, and thus the low-density portion is also enhanced by the subsequent frequency emphasis also the noise is amplified. Thus, it is preferable that each signal is treated on the basis of the original image signal S _org . From the viewpoint of saving processing time, it is also preferable if both processings are performed in parallel.

By calculating the transform functions such that the frequency response of the processed image signal S _proc becomes constant independently of the resolution of the original image signal S _org , the processed image signal S _{proc may have} a substantially constant frequency response independent of the resolution of the original image signal S _org .

Although in the third to fifth embodiments, the transformation functions are calculated by correcting the reference transformation functions based on the input resolution of the original image signal S _org , the obtained transformation functions can be further corrected based on information of the response characteristic of the original image signal S _org . That is, the original image _signal S _org is obtained from various devices, for example, a semiconductor sensor other than the radiation image reader, when a radiation image is read out on a stimulable phosphor sheet, the response of the original image _signal S _{org being} different from device to device. By correcting the transformation functions on the basis of the information about the response characteristic of the original image signal S _org , finally, a processed image signal S _{proc can be} obtained which has a frequency response independent from the respective device from which the original image signal S _{org was} obtained. For example, if the response at 2 cycles / mm of an original image signal S _{org received} from a particular device is 40% below a reference response, the circumstance is input to the image processing system and this corrects the transform functions such that the response for 2 cycles / mm of the reference response.

Although at the fourth and fifth, embodiment described above non-linear functions are used and regarding the band-limited signals a non-linear processing is performed can as transformation functions also linear functions or constants be used.

In the third to fifth embodiments described above, although the blur image signals are obtained from the original image signal by filtering and interpolating / enlarging, and the band-limited signals are obtained from the original image signal and the blur image signals, the band-limited signals may be formed by transforming the original image signal into multi-resolution image signals, for example be generated by means of a wavelet transformation or a Laplace pyramid, wherein the blur image signals from the image signals with the respective triggers generated and the band-limited th signals are obtained from the blur image signals.

When the transform functions for transforming band-limited signals obtained from an original image signal different in resolution from the reference original image signal are obtained by correcting the reference transform functions, it is generally not possible that the frequency response of the processed image signal S _proc with the processed reference image signal S _proc 0 (this is the processed image signal S _proc obtained by transforming band-limited signals from the reference image signal using the reference transform functions) in the entire frequency bands. In an image obtained by reproducing a processed image signal, the low-frequency components are visually more discernible than the high-frequency components. Accordingly, it is preferable that the band-limited signal transforming functions for the original image signal to be processed are calculated such that the frequency response of the processed image signal S _proc coincides with that of the processed reference image signal S _proc 0 within frequency bands of not less than 1/5, More preferably, not less than 1/2 of the Nyquist frequency of the original image signal to be processed. For example, if the resolution of 5 lines / is mm and the resolution of the processed original image signal S _org 4 lines / is mm, so it is preferred that the transformation functions for transforming the extracted from the original image signal S _org band-limited signals on the basis of reference Transformation functions for 5 lines / mm can be calculated in the following way.

When the image signal is gradually reduced by 1/2 and a plurality of band-limited signals are obtained, the band-limited signal in the highest frequency band has at least a characteristic width of 1.0 to 2.0 cycles / mm. Consequently, in the image processing system according to the third to fifth embodiments of the invention, in which the gains of the band-limited signals are controlled by the transform functions, it is impossible to finely control the response in the frequency band between 1.0 and 2.0 cycles / mm. and thus, if the frequency response of a sharp change in the frequency band is between 1.0 and 2.0 cycles / mm, it is impossible to obtain the transform functions for the 4 lines / mm, the precise match with the frequency response of the processed signal S _proc and bring the one of the processed reference _image signal S _proc 0 in the frequency band between 1.0 and 2.0 cycles / mm. Consequently, it is preferable that the transform functions for band-limited signals of the original image signal (4 lines / mm) be calculated such that the frequency response of the processed image signal S _proc coincides with that of the processed reference image signal S _proc 0 within those frequency bands not lower than 1 / 5, more preferably not less than 1/2 of the Nyquist frequency of the original image to be processed, as in 39A and 39B is shown. In 39A That is, the broken line indicates the frequency characteristic of the processed reference signal S _proc 0 obtained by transforming the band-limited signals from the reference image signal (with a pixel density of 5 lines / mm) using reference transformation functions set to amplify relatively low-frequency components while the solid line shows the frequency response of the processed image signal S _proc obtained by transforming the band-limited signals from the original image signal coffin to be processed by using the transform functions obtained by correcting the reference transform functions according to the resolution (4 lines / mm ) of the original image _signal S _org to be processed. In 39B the dashed line shows the frequency response of the processed reference _image signal S _proc 0 obtained by transforming the band-limited signals from the reference image signal (at a resolution of 5 lines / mm) by using reference transformation functions set to amplify relatively high frequency components while the solid line represents the frequency response of the processed image signal S _proc obtained by transforming the band-limited signals from the original processing image signal S _org using transformation functions obtained by correcting the reference transformation functions according to the resolution (4 lines / mm ) of the original image _signal S _org to be processed. An image processing system according to a sixth embodiment of the invention will be explained below. The image processing system is to perform frequency emphasis by using blur image signals of an original image signal obtained by reading a radiation image of a human body recorded on a stimulable phosphor sheet to obtain an image suitable for diagnosis. An image reproduced on the basis of the processed image signal is mainly recorded on a photographic film and used in the diagnosis.

In 40 contains an image processing system 201 according to a sixth embodiment of the invention, a blur image signal generating means 202 , The multiple resolution image signals S _org, which are obtained from an original image signal S _org by transforming this signal into a multi-resolution space and encoding the transformed image signals' from a plurality of unsharp image signals he and a frequency emphasis processing device 203 which performs frequency emphasis to emphasize a particular frequency and obtains a processed intermediate image signal S _proc '. The image processing system 201 also contains a parameter setting device 204 , a scale input device 205 and an enlargement / contraction device 206 , The parameter setting device 204 is a device that sets transformation functions that the frequency emphasis processing device 203 used in the frequency emphasis, based on scale information S, via the scale input device 205 is entered. The scale input device 205 is means for obtaining scale or scale information S with respect to the original image _signal S _org . This scale information S may be entered by the operator via a keyboard as a value, or by the operator selecting a value from a plurality of scales displayed on a control screen. The enlargement / contraction device 206 Enlarges or contracts a processed intermediate _image signal S _proc ', which of the frequency emphasis processing means 203 on the basis of the scale information S, thereby obtaining a final processed image signal S _proc .

The original image signal S _org is transformed into a multi-resolution area, and the transformed image signals are encoded in the following manner. As in 41A 2, the original image signal S _{org is} subjected to a wavelet transformation and decomposed into four pieces of data or blocks LL1, HL0, LH0 and HH0 in a resolution-dependent manner. Data LL1 represents an image obtained by reducing the original image 1/2 in both the longitudinal and lateral directions. The data HL0, LH0 and HH0 represent a longitudinal edge, a side edge and an oblique edge. When the four data pieces LL1, HL0, LH0 and HH0 are subjected to an inverse wavelet transformation, the original image signal S _org is again obtained. Then, the data LL1 is further subjected to wavelet transformation and decomposed into the four pieces of data LL2, HL1, LH1 and HH1, as in FIG 41B you can see. The data LL2 represents an image obtained by reducing the data LL1 by 1/2 in both the longitudinal and lateral directions. The data HL1, LH1 and HH1 represent a longitudinal edge component, a side edge component and a skew edge component of the data LL1. When the four data pieces LL2, HL1, LH1 and HH1 are subjected to inverse wavelet transformation, a low resolution image signal having a resolution equal to 1/2 of that of the original image signal is obtained. The wavelet transform is repeated at a desired frequency with respect to the data LL obtained every time the wavelet transform is performed, so that finally several pieces of data having different resolutions are obtained. Subsequently, the data for each resolution according to 41C coded, and you get image signals with multiple resolution, S _org '.

By decoding only the image signals up to a desired resolution from the multi-resolution image signals S _org 'and performing the inverse wavelet transformation on the image signals, a low-resolution image signal can be obtained which is an image having a resolution of 1/2 ^k (k corresponds to the desired resolution) of that of the original image.

The generation of the blur image signals will now be described. As in 42 is shown, contains the blur image signal generating means 202 a first to nth filter device 210 , The first filter device 210 performs filter processing on the original image signal S _org reconstructed from the multi-resolution image signals S _org in x- and y-directions, and forms a low-resolution image signal B1, which has a lower resolution than the original image signal S _org . The second filter device 210 performs similar filtering processing on the low-resolution picture signal B1, and generates a low-resolution picture signal B2, in which the resolution is lower than that of the picture signal B1, and the third filter means 210 performs similar filtering processing on the thus obtained low resolution image signal B2 to produce a low resolution image signal B3 whose resolution is lower than that of the image signal B2. In this way, the nth filter device performs 210 performs similar filtering of the low resolution image signal Bn-1 and generates a low resolution image signal Bn whose resolution is smaller than that of the image signal Bn-1. The blur image signal generator 202 also contains a first to nth interpolation device 211 , The interpolation device 211 performs interpolation / enlargement processing on the image signals B1 through Bn of low resolution obtained by the individual filter stages, and generates a plurality of unsharp image signals S _usk (S _us 1 through S _us N) having different sharpness.

at this particular embodiment become filters with essentially a one-dimensional Gaussian distribution for the Filtering used. This means: Filter coefficients are determined according to the following Formula (4), which refers to a Gaussian signal.

This is because the Gaussian signal provides good order in both real space and frequency space. For example, the in 43 used one-dimensional 5 × 1 filter, where σ = 1 in the formula (4).

The filtering is performed with respect to the original image signal S _org or the low resolution image signals every other picture element as in FIG 44 you can see. When such filtering is performed in both the x and y directions, the number of pixels in the obtained low resolution image signal reduces to 1/4 of the previous image signal (for example, the original image signal S _org less in the case of the image signal B1 Resolution, and the low resolution image signal B1 in the case of the low resolution image signal B2). That is, the number of picture elements in each of the low-resolution picture signals Bk (k stands for 1 to n) is 1/2 ^{k of} the original picture signal S _org .

The interpolation enlargement processing to be performed with respect to the thus-obtained low-resolution image signals Bk will be described below. Although various methods of interpolation, such as the B-spline method, may be used, a Gaussian signal is used here for interpolation because the low-pass filters based on Gaussian signals are used for filter processing. Specifically, in the following formula (5), an approximation σ = 2 ^{k-1 is} used:

When interpolating the image signal of low resolution, B1, because of k = 1, σ = 1 holds. In this case, an in 45 illustrated one-dimensional 5 × 1 filter used for the interpolation. In this interpolation, a picture element having the value 0 is interpolated into every other picture element of the low-resolution picture signal B1, this picture signal B1 extended to a size corresponding to the original picture size, and the low-resolution interpolated picture signal B1, by means of the in 45 is subjected to a filtering shown one-dimensional filter.

This interpolation-enlargement processing is performed with respect to all the low-resolution image signals Bk (B1 to Bn). In interpolating the low-resolution image signal Bk, a filter having a length of 3 × 2 ^k -1 corresponding to the formula (5) is prepared, and 2 ^k -1 picture elements having the value 0 are inserted between each pair of adjacent pixels, whereby the image signal Bk low resolution is widened to a size that corresponds to the original image size. Then, the interpolated Bildsi signal Bk low resolution using the filter of length 3 × 2 ^k -siluted filtering.

In the following, the frequency emphasis performed on the blur image signals S _usk thus obtained will be described. 46 shows a system for performing a frequency emphasis together with the blur image signal generating means 202 , As 46 shows, unsharp image signals S _us k from the original image signal S _org formed, and there are differences between the original image signal S _org and the unsharp image signals S _us k by respective subtractors 221 formed, whereby band-limited signals (S _org - S _us 1, S _us 1 - S _us 2 and the like) arise whose components are in the restricted frequency bands of the original image signal S _org .

The band-limited signals are adjusted to predetermined amplitudes by means of different transformation functions f ₁ to f _N in the associated transformation circuits 222 transformed, and the transformed band-limited signals are from an operator 223 is added to the original image signal S _org according to the formula (2 ') given below, whereby the processed intermediate image signal S _proc ' is generated, in which a desired frequency component is amplified to a required degree. S proc '= S org + β (S. org ) × F usm (S org , P us 1, p us 2, ..... S us N) F usm (S org , P us , P us 2, ..... S us N) = f 1 (S org - p us 1) + f 1 (S us 1 - p us 2) + ... + f k (S us k - 1 - p us k) + ... + f N (S us N - 1 - S us N) (2 ') where S _proc 'is a processed intermediate image _signal in which the high frequency components are emphasized, S _{org is} an original image _signal , S _usk (k = 1 to N) are unsharp image signals, f _k (k = 1 to N) is a transform function, and β (S _org ) is an emphasis coefficient that is determined based on the original image signal becomes.

In this way, the processed intermediate _{image signal} S _proc 'is created. The problem to be solved by this embodiment and the manner of solving the problem according to this embodiment will be explained below by way of example. As described above, although a processed intermediate image _signal S _proc 'corresponding to an image equivalent to the original image signal can be obtained in its resolution (either the same image as the original image or a different image) by emphasizing a frequency using transformation functions f _k However, a processed intermediate image signal S _proc ', which represents an image having a lower resolution than the original image signal (either the same image as the original image or a different image), must be obtained by: _{## EQU1 ##} which is determined according to the frequency bands of the band-limited signals frequency emphasis using transformation functions f _k determined according to the resolution of the processed intermediate _image signal S _proc . If such transformation functions are stored for the individual resolutions, the number of transformation functions to be stored becomes too large and the handling of these transformation functions becomes too cumbersome. When the frequency _{emphasis processing is} performed on an original image signal S _org when it is separated into six frequency bands, the frequency response of the band-limited signals has six peaks, as in FIG 47 is shown. The frequencies of the individual peaks are 5 cycles / mm; 1.0 cycle / mm; 0.5 cycles / mm; 0.25 cycles / mm; 0.12 cycles / mm or 0.06 cycles / mm. In an image signal of an image having a resolution of 1/2 of that of the original image signal, the peak frequencies of the band-limited signals are 2.5 cycles / mm; 0.5 cycles / mm; 0.25 cycles / mm; 0.125 cycles / mm and 0.06 cycles / mm, as in 48 you can see. As continue in the 49 to 51 is shown, in image signals of images whose resolutions are 1/4; 1/8 and 1/16 of that of the original image, the peak frequencies of the band-limited signals are equal to each other except their uppermost frequency bands.

In this embodiment, when a processed image signal corresponding to an image having a desired resolution lower than the resolution of the original image is to be obtained, one of the low-resolution signals corresponding to a low-resolution image becomes closest to that having the desired resolution is used as a low-resolution reference image signal, and based on this signal, low-resolution band-limited signals are obtained and transformed into image signals based on transform functions corresponding to the frequency bands of the band-limited signals, thereby obtaining a processed intermediate image signal S _proc '. Then, the processed intermediate image signal S _proc 'is enlarged or contracted depending on the scale information S, and a processed image signal S _proc corresponding to an image having the desired resolution is obtained.

With reference to the in 52 In the flow chart shown below, the operation of the image processing system 201 of this embodiment will be described below when a processed image signal S _proc corresponding to an image having a resolution of 1/2 that of the original image signal is to be obtained. It is assumed that the band-limited signals obtained from the original image signal S _org are in six frequency bands which are in 47 is shown.

First, the user gives via the scale input device 205 a desired scale (step S21). Then the scale input device gives 205 in the blur image signal generation means 202 , the parameter setting device 204 and the enlargement / contraction device 206 a scale information S in, and the blur image signal generating means 202 reconstructs a low-resolution image signal representing an image having a resolution corresponding to that of the desired scale from the low-resolution image signals B _k (S22) based on the scale information S, and produces blur image signals S _us k less Resolution from the image signals of low resolution (step S23). In this example, since the resolution of the image represented by the processed image signal S _{proc to be obtained} (herein referred to as "the desired" or "target resolution") is 1/2 of that of the image represented by the original image signal S _org , blur image signals S _us 1 to S _us 5 in correspondence between the blur image signals obtained from the original image signal S _org (including the original image signal S _org ) and those obtained from a low resolution image signal, is shown in FIG 53 shown. How out 53 is apparent, when the desired resolution 1/2 corresponds to that of the original image, blur image signals S _us k are prepared such that the blur image signal S _us 1 which is in the highest hierarchical level of the blur image signal obtained from the reference original image signal S _org is, as original _{image signal} S _{org1 / 2} . Consequently, the original image signal S in this case corresponds _{org1 / 2} the low resolution obtained from the reference original image signal S _org unsharp image signal S _us 1, the unsharp image signals S _us 1 _1/2 corresponds to the image signals S _us 2 obtained from the reference original image signal S _org , the blur image signals S _us 2 _1/2 correspond to the blur image signals S _us 3 obtained from the reference original image signal S _org , etc. Subsequently, the band-limited signals become low resolution on the basis of the original _{image signal} S _{org1 / 2} and the low resolution blur _{image signals} S _us 1 _1/2 to S _us 5 _1/2 (step S24). The correspondence between the band-limited signals obtained from the reference original image signal S _org and those obtained from an image signal of a given low resolution is shown in FIG 54 shown. Even if the band-limited signals are processed by not more than 0.06 cycles / mm, the processed image signal S _proc does not appreciably improve in quality, the low frequency band-limited signals of not more than 0.06 cycles / mm are not processed.

The parameter setting device 204 sets the transformation functions f _k for transforming the low-resolution band-limited signals (step S25). Since the peak frequencies of the low-resolution band-limited signals are equal to those of the band-limited signals obtained from the reference original image signal S _org , except for their highest frequency bands as in Figs 47 and 48 is shown, the transformation functions f ₁ to f ₆ are used. The correspondence between the transformation functions for the band-limited signals obtained from the reference original image signal S _org and those for the band-limited signals obtained from a given low-resolution image signal is shown in FIG 55 shown. If the desired resolution is equal to 1/2 ^{n of} the resolution of the reference original image signal, the transform functions in that line in 55 verwen det, which is n lines below the transformation functions that apply to the reference original image signal S _org . The following formula (14) provides Fusm in the aforementioned formula (2 ') in performing the frequency emphasis on the reference original image signal S _org , and the following formula (15) for fusm in the above formula (2') stands for performance the frequency _emphasis with respect to the original _{image signal} S _{org1 / 2} with a resolution of 1/2 of that of the reference original image signal S _org . F usm (S org , P us 1, p us 2, ..... S us 6) = f 1 (S org - p us 1) + f 2 (S us 1 - p us 2) + ... + f 5 (S us 4 - p us 5) + f 6 (S us 5 - p us 6) (14) F usm (S org1 / 2 , P us 1 1.2 , P us 2 1.2 ..... S us 5 1.2 ) = f 1 (S org1 / 2 - p us 1 1.2 ) + f 2 (S us 1 1.2 - p us 2 1.2 ) + ... + f 4 (S us 3 - p us 4) + f 5 (S us 4 - p us 5) (15)

Then, low-resolution transformed image signals are generated from the band-limited low-resolution signals (step S26) to obtain a low-resolution addition signal. Then, the low-resolution addition signal is multiplied by the emphasis coefficient β (S _{org1 / 2} ), and the product is added to the original _{image signal} S _{org1 / 2 of} low resolution, thereby _{obtaining a} low-resolution processed intermediate _{image signal} S _{proc1 / 2} '. The stress _coefficient β (S _{org1 / 2} ) is set according to the original _{image signal} S _{org1 / 2 of} low resolution.

The processed low resolution intermediate _{image signal} S _{proc1 / 2} 'thus obtained is _input to the enlargement / contraction means 206 _inputting or contracting the image signal S _{proc1 / 2} 'on the basis of the scale _information S _input via the scale input means 205 so as to be reproducible in the low-resolution image with the desired scale to thereby obtain a final processed low-resolution image signal S _{proc1 / 2} to get. In this particular example, since the desired resolution is 1/2, the magnification / contraction device needs to be increased 206 the processed intermediate _{image signal} S _{proc1 / 2} 'not to increase or contract.

As apparent from the above description, the low-resolution band-limited signals obtained from the original low-resolution image signal are transformed based on a part of the transform functions for transforming the band-limited signals obtained from the reference original image signal corresponding to the band-limited ones Therefore, the image signal reproduced on the basis of the processed low-resolution image signal S _{proc1 / 2} in the frequency response substantially matches the image reproduced on the basis of the processed image signal S _proc obtained from the reference original image signal whereby a processed image signal can be obtained which can reproduce an image having a substantially constant frequency response regardless of the resolution.

In addition, since some of the transformation functions used to emphasize the frequency of the original image signal are used, transformation functions do not need to be created for each resolution which simplifies the arrangement and design of the image processing system and avoids difficulties in handling a large number of transformation functions.

When an image signal having a desired resolution is obtained by contracting a processed image signal S _proc which has been obtained from the reference original image signal S _org in the generation of blur image signals, the generation of band-limited signals as well as the transformation of the band-limited signals must be performed all image signals of low resolution are performed. In contrast, in this embodiment, the processing time required for processing can be shortened since only the low-resolution image signals up to those representing an image having a resolution closest to the desired resolution need to be reconstructed.

Although in the sixth embodiment described above, the blur image signals from the original image signal (the reference original image signal or the original image signal of low resolution) by filtering and interpolation / enlargement and the band-limited signals from the original image signal and the blur image signals can be won For example, the band-limited signals are generated by transforming of the original image signal in multi-resolution image signals by means of a wavelet transform or a Laplace pyramid by the blur image signals be generated from the image signals with the respective resolutions and the band-limited signals are in turn generated from the blur image signals.

Although the frequency emphasis processing with respect to encoded multiple resolution image signals S _org conducted at the above-described sixth embodiment ', the frequency enhancement processing can be carried out also with respect to the non-coded original image signal S _org.

An image processing system according to a seventh embodiment of the invention will be described below. In the sixth embodiment, low-resolution image signals up to the closest resolution to a desired resolution are reconstructed from coded multi-resolution image signals S _org ', and blur image signals are generated based on the reconstructed low-resolution image signals. In contrast, in the seventh embodiment, blur image signals are generated on the basis of low-resolution image signals ranging from the lowest resolution to the closest resolution to the resolution corresponding to a desired scale. The seventh embodiment differs from the sixth embodiment only in the processing performed by the blur image signal generation means, so that the processing performed by the blur image signal generation means in this embodiment will be mainly described below.

56 shows a blur image signal generator 202 within the image processing system of the seventh embodiment. As in 56 is shown, contains the blur image signal generating means 202 ' a plurality of encoders 230 and a plurality of interpolators 231 , That is, a low-resolution image signal B _n corresponding to an image having a resolution of 1/2 ^{n of} that of the original image represented by the reference original image signal S _org is first formed by decoding an image signal representing an image among the images corresponding to the multi-resolution image signals S _org 'has the lowest resolution, and by performing an inverse wavelet transformation on the decoded image signal. Subsequently, a low resolution image signal B _n-1 corresponding to an image having a resolution of 1/2 ^{n-1 of} that of the original image is formed on the basis of the image signal B _n . In this way, image signals of low resolution, B _k , whose resolution 1/2 ^k (k = 1 to n) of that of the original image signal S _org as well as the original image signal S _{org are} generated. The low-resolution image signals B _k correspond to the data LLk obtained by performing wavelet transformation on the original image signal S _org . Then the interpolation device leads 231 interpolation / enlargement with respect to each of the low-resolution image signals B _k , and a plurality of unsharp image signals S _us k (k = 1 to N) are formed which are different in sharpness from each other.

Then band-limited signals are obtained from the blur image signals S _usk , and according to the above formula (2 '), a processed intermediate image signal S _proc ' is obtained.

By way of example, the operation of the image processing system of this embodiment will be described in the case that a processed image signal S _proc is to be obtained, which represents an image whose resolution 1/2 corresponds to that of the original image signal. It is assumed that the band-limited signals obtained from the original image signal S _org are in six frequency bands.

First, the user gives via the scale input device 205 a desired scale. The scale input device 205 then gives scale information S to the blur image signal generation means 202 ' , the parameter setting device 204 and the enlargement / contraction device 206 on, and the blur image signal generator 202 ' reconstructs low-resolution image signals B _k from the lowest resolution up to the resolution nearest the resolution corresponding to the desired scale based on the scale information S, and generates low-resolution blur image signals S _us k. In this example, since the resolution of the image represented by the processed image signal S _{proc to be obtained} (hereinafter also referred to as the desired or "target resolution") is 1/2 of that of the image represented by the original image signal S _org , blur image signals S _us 1 to S _us 5. The correspondence between the blur image signals obtained from the original image signal S _org (including the original image signal S _org itself) and those obtained from the low resolution image signals B _k from the lowest resolution to one of the desired scale corresponding resolution next resolution, is in 53 shown. The correspondence between the band-limited image signals obtained from the reference original image signal S _org and those obtained from the low-resolution image signals B _k from the lowest resolution to a resolution which corresponds to the resolution nearest to the desired scale is in 54 shown.

The parameter setting device 204 sets the transformation functions f _k for transforming the low-resolution band-limited signals in the same manner as in the sixth embodiment. If the desired resolution 1 / 2n matches the resolution of the reference original image signal, the transform functions are used in the line written in 55 by n lines below the transformation functions for the reference original image signal S _org .

Then, low-resolution image-transformed image signals are generated from the band-limited low-resolution signals, and a low-resolution addition signal is formed. This low-resolution addition _signal is then multiplied by the emphasis _coefficient β (S _{org1 / 2} ), and the product is added to the original low-resolution _{image signal} S _{org1 / 2} , thereby obtaining a low-resolution processed intermediate _{image signal} S _{proc1 / 2} '.

The thus processed processed intermediate _{image signal} low resolution S _{proc1 / 2} 'is in the enlargement / contraction _device 206 input the signal S _{proc1 / 2} 'on the basis of the scale _{input device} 205 input scale information S is increased or contracted so that a low-resolution image can be reproduced according to the desired scale to obtain a final processed low-resolution image signal S _{proc1 / 2} . Since the desired resolution is 1/2 in this particular example, the device needs 206 not to increase or contract the signal S _{proc1 / 2} '.

As can be understood from the above description, also in this embodiment, the low-resolution band-limited signals obtained from the original low-resolution image signal are converted based on a part of the transform functions for transforming the band-limited signals obtained from the reference original image signal into the Frequency band corresponding to the low-resolution band-limited signals, and hence the image reproduced on the basis of the processed low-resolution image signal S _{proc1 / 2} corresponds to the image reproduced on the basis of the processed image S _proc obtained from the reference frame. Original image signal, so that a processed image signal can be obtained, which can reproduce an image, which has a substantially constant frequency response, regardless of the resolution.

Although the above description relates to an example of a case where the desired resolution is 1/2 that of the reference original image signal, at a resolution of, for example, 1/3 of that of the reference original image signal, the blur image signal generating means becomes 202 and the frequency emphasis processing means 203 the frequency emphasis with respect to the image signal B _{2 of} low resolution corresponding to an image with a resolution (1/4 of the original image signal) closest to 1/3, so as to come to a processed intermediate _{image signal} S _{proc1 / 4} '. The final processed image signal S _{proc1 / 4} is then achieved by having the processed intermediate image signal S _{proc1 / 4} ' _4/3 times from the enlargement / contraction means 206 is enlarged. If the desired resolution is 1/2 _n that of the original image, the enlargement / contraction device 206 be omitted, which simplifies the system.

Although in the sixth to seventh embodiments, the transformed image signals thereby If the band-limited signals are transformed on the basis of the transform functions f _k according to the above formula (2 '), the transformed signals can also be obtained by using the band-limited signals by means of emphasis _coefficients α _k (k stands for 1 to N) instead of the transformation functions f _k as shown by the following formula (16): S proc '= S org + β (S. org ) × F usm (S org , P us 1, p us 2, ..... S us N) F usm (S org , P us 1, p us 2, ..... S us N) = α 1 (S org - p us 1) + α 2 (S us 1 - p us 2) + ... + α k (s us k - 1 -S us k) + ... + α N (S us N - 1 - S us N) (16)

When an original image signal to be processed has a higher resolution than the reference original image signal S _org , the number of band-limited signals and the number of transformation functions for transforming them are larger than those of the reference original image signal S _org . 59 _{Fig. 15} shows the correspondence between the band-limited signals obtained from the reference original image signal S _org and those from the original image signals whose resolution is twice and four times that of the reference original image signal S _org . If the object original image signal to be processed has twice the resolution as the reference original image signal S _org (this object original image signal is written in 57 designated S _org 2, and the blur _{image signals} obtained therefrom are designated S _us2 k (k stands for an integer greater than 0)), so there is no band-limited signal within the band-limited signals obtained from the reference original image signal S _org corresponding to band-adjacent signal S _org2 - S _us2 1 in the highest frequency band de from the original object image _signal S _org2 obtained band-limited signals, and accordingly there is no transformation function for transforming the band-limited signal S _org2 - S _us2 1 in the reference transformation functions for transforming the reference original image signal S _org obtained, band-limited signals. Similarly, if the object original image signal to be processed has a resolution four times as large as the reference original image signal S _org (this object original image signal is written in _FIG 57 with S _org4 , and the blur _{image signals} obtained therefrom are _denoted by S _us4 k (k is an integer greater than 0)), there is no band-limited signal within the band-limited signals obtained from the reference original image signal S _org corresponding to band-limited signals S _org4 - S _us4 1 and S _us4 1 - S _us4 2 in the highest and second highest frequency bands within the band-limited signals obtained from the original object image _signal S _org4 , and accordingly, there is no transformation function for transforming the band-limited signals S _org4 - S _us4 1 and S _us4 1 - S _us4 2 among the reference transformation functions for transforming the band-limited signals obtained from the reference original image signal S _org . In such cases, the transform function f1 for the band-limited signal S _org -S _us 1 in the highest frequency band among the band-limited signals from the reference original picture signal S _{org is used} as a transform function for transforming the band-limited signal (s) in the highest and second highest frequency band derived from Object original image signal was obtained as in 58 is shown.

When a plurality of reference original image _signals S _{org have been} processed, it is preferable that one of the reference original image _signals S _org closest to the object original image signal in the resolution be used as the reference original image signal S _org in the above in connection with _FIG in the first and second embodiments. It is preferable that the transformation functions determined for the band-limited signals obtained from the object original image signal are related to the object original image signal and stored together with the resolution of the reference original image signal. In this embodiment, the image represented by the object original image signal having the same characteristics can be reproduced in a constant manner.

in the The following will be an image processing system according to a seventh embodiment described. The image processing system is used to perform a Frequency emphasis using blur image signals of an original image signal, which by reading one on a stimulable phosphor sheet recorded radiation image of a human body was, so that one for one Diagnosis suitable image can be obtained. A based on The image reproduced from the processed image signal becomes the first Line recorded on a photographic film and at the Diagnosis used.

In 59 contains an image processing system 301 according to the seventh embodiment of the invention, a pixel density transformation means 304 , a blur image signal generator 302 a frequency emphasis processing means 303 , an enlargement / contraction device 305 and a parameter corrector 306 , When a frequency emphasis is to be made with respect to an original image signal S _org itself, the pixel density transformation means leaves 304 the original image signal S _org unchanged from the blur image signal generator 302 by generating a plurality of blur image _signals from the original image signal S _org and inputting them to the frequency emphasizing processing means 303 performs a frequency emphasis to emphasize a particular frequency to obtain an intermediate processed _image signal S _proc 'which is input to the enlargement / contraction means 305 is input, the unchanged then outputs the intermediate processed image signal S _proc 'as a final processed image signal S _proc. On the other hand, when the frequency _{emphasis is to be} performed with respect to an original image signal S _org obtained by pixel density transformation of a basic original image signal or basic image signal S _org 'read by an image reader or the like and a predetermined reference pixel density (in this particular Embodiment, 10 lines / mm), the picture element density transform means transforms 304 the pixel density of the basic original image signal S _org 'to a target pixel density of, for example, 6.7 or 5 lines / mm to thereby obtain an original image _signal S _org having the desired or desired pixel density. Subsequently, the pixel density transforming means outputs 304 the picture element density-transformed original picture signal S _org into the blur picture signal generating means 302 which generates a plurality of blur image signals from the original image signal S _org and _input them into the frequency emphasis processing means 303 enters. This performs the frequency emphasis to emphasize a particular frequency to obtain a processed intermediate _image signal S _proc 'which is input to the enlargement / contraction means 305 is entered. The enlargement / contraction device 305 enlarges the intermediate processed image signal S _proc 'by interpolation of a desired magnification and outputs the enlarged intermediate processed image signal S _proc' as a final processed image signal S _proc. In the present specification, the term "magnification" is to be broadly interpreted, that is also to include a magnification of less than 1. The desired pixel density may be entered as a value via a keyboard or the like, or it may be selected from a set of pixel densities Similarly, the desired magnification may be entered as a value via a keyboard or the like, or may be selected from a plurality of magnifications displayed on a control screen.

The pixel density transformation processing will be described below. In the above-explained radiation image reading system using a stimulable phosphor sheet, the reading density or the pixel density differs depending on the size of the stimulable phosphor sheet and can be freely changed as desired by the user. In the present embodiment, a basic original image _signal S _org 'read at 10 lines / mm is transformed into an original image _signal S _org at 6.7 or 5 lines / mm by converting the basic original image signal S _org ' from the pixel density transform means 304 subjected to a filter processing. The original image signal S _org obtained by carrying out the pixel density transformation with respect to the original original image signal S _org is _hereinafter referred to as "original image signal S _org " transformed into the pixel _density , to distinguish from the former signal.

The blur image signal generator 302 generates blur image signals S _us k in the same manner as explained above and the frequency emphasis processing means 303 performs the frequency emphasis using the blur _image signals S _usk in the manner described, so that a processed intermediate _{image signal} S _proc 'is obtained.

The enlargement / contraction device 305 enlarges the intermediate processed image signal S _proc 'by interpolation using an interpolation filter at a desired magnification, thereby obtaining a final processed image signal S _proc.

Desired frequency components of the original image signal S _org can be emphasized by thus performing frequency emphasis on the original image signal S _org by using transforming functions arranged for the respective frequency bands of the band-limited signals. However, when the frequency emphasis on original image signals S _org transformed in the pixel _{density is} performed by using the same transformation functions f _k as those used in performing the frequency emphasis on the basic original image signal S _org ', this differs on the basis of FIG the processed image signal extracted in the pixel density S _org reproduced image in the frequency response of that which is reproduced on the basis of the processed image signal S _proc obtained from the basic original image signal S _org ', and the first image is less sharp than the latter although the pictures are the same size. This is due to the pixel density transformation performed when the original image signal S _org transformed in the pixel _{density is} obtained and the interpolation processing is performed to enlarge the processed intermediate image signal S _proc '.

In this embodiment, the transform functions f _k for transforming the band-limited signals obtained from the original image signal S _org transformed in the pixel density are obtained by using the parameters of the transform functions f _k for transforming the band-limited signals obtained from the original original image signal S _org according to the filter characteristic F1 of the pixel density transformation filter used for the pixel density transformation with the pixel density transformation means 304 is corrected as well as the filter characteristic F2 of an interpolation filter for performing the interpolation using the enlargement / contraction means 305 such that the frequency response of the image reproduced on the basis of the processed image signal S _{proc obtained} from the original image signal S _org obtained in the pixel density is equal to the frequency response of the image formed on the basis of the original original image signal S _org '. obtained processed image signal S _proc .

The following is the correction of the parameters for the transformation functions by the parameter correction device 306 be explained.

60 FIG. 11 shows the filter characteristic F1 of the pixel density transformation filter obtained by the pixel density transformation means 304 is used to the base original image signal S _org 'having a picture element density of 10 lines / mm to convert into a transformed to the picture element density original image signal S _org, which / has a picture element density of 5 lines mm, and 61 shows the filter characteristic F2 of the interpolation filter for the enlargement / contraction device 305 for doubling the original image signal S _org transformed in the pixel _density . The latter signal S _org obtained by transforming the base original image signal S _org 'having a picture element density of 10 lines / mm using the picture element density-transformation filter having the filter characteristic F1 in the transformed in the pel density original image signal S _org having a picture element density of 5 lines / mm, is hereinafter referred to as "5-line transformed original image signal S _org 5". 62 shows the frequency response of the processed image signal S _proc , which is obtained by performing the frequency emphasis using the transformation functions f _k , for the frequency emphasis of the original original image signal S _org '(these transformation functions are hereinafter referred to as "basic transformation functions f _kb ". with respect to the 5-line transformed original _{image signal} S _org5 and by doubling the obtained processed intermediate _{image signal} S _proc 'using the interpolation filter having the filter characteristic F2. The processed image signal S _proc obtained in this way from the 5-line-transformed original image signal S _org 5 is referred to below as "5-line processed image signal S _proc 5." Furthermore, this is emphasized by emphasizing the frequency of the basic original image signal S _org 'using the basic image signal S _proc . Transformation functions f _kb received processed image signal S _proc referred to as "the processed basic image signal S _Proc B". In 62 the frequency response of the 5-line processed image signal S _Proc 5 is shown in conjunction with the frequency response of the processed basic image signal S _Proc B with the size of 1. How out 62 can be removed, the frequency response of the 5-line processed image signal S _proc 5 can be calculated in a simple manner on the basis of the filter characteristic F1 of the pixel density transformation filter and the filter characteristic F2 of the interpolation filter. It also turns off 62 in a simple manner, the difference in frequency response between the base processed image signal S _proc B and the 5-line processed image signal S _proc 5 win.

To _proc for example, the resolution ratio of, on the basis of the 5-line processed image signal S adapt 5 reproduced image to the conditions of on the basis of the base processed image signal S _proc B at 2 cycles / mm reproduced image, it is necessary, the frequency response at 2 cycles / mm of the 5-line processed image signal S _proc 5 to double. Thus, in this embodiment, the band-limited signal transform function in the 2-band / mm frequency band is adjusted by correcting the parameters for the basic transform function f _kb for the band-limited signal in the frequency band corresponding to 2 cycles / mm 2 cycles / mm and thus the stress level is doubled. As in 63 2, the band-limited signal is located in the frequency band corresponding to 2 cycles / mm in the highest frequency band, and hence the parameter of the band-limited signal transformation function f ₁ in the highest frequency band is corrected so as to double the amount of emphasis or gain. Specifically, if the transform function is a slope function, the slope is doubled, and if the transform function is a constant, its value is doubled. The frequency component at which the resolution ratio of the image 5 reproduced on the basis of the 5-line image signal S _proc 5 is adapted to that of the image reproduced based on the processed basic image signal S _proc B any frequency, which is also two or more Frequencies can act. Furthermore, the frequency component can be input by the user via an input device, not shown.

Hereinafter, the operation of the image processing of this embodiment will be described with reference to FIG 64 illustrated flowchart described. It is assumed that a 5-line processed image signal S _proc 5 is obtained by transforming a basic original image signal S _org 'with a pixel density of 10 lines / mm into a 5-line processed original image signal S _org 5 with a pixel density of 5 lines / mm, carrying out the frequency enhancement processing on the 5-line transformed original image signal S _org 5, thereby an intermediate processed image signal S _proc to be obtained, and doubling the intermediate processed image signal S _proc '. The basic original image _signal S _org is first _input to the image processing system 301 entered (step S31). Then, the pixel density transforming means transforms 304 the basic original image signal S _org 'into the 5-line original image signal S _org 5 (step S32). This is sent to the blur image signal generator 302 entered, the unsharp image signals S _us k from the 5-line transformed original image signal S _org 5 forms (step S33). The filter characteristics F1 and F2 are put into the parameter correction device 306 entered. Further information of the frequency component with respect to the adjust the resolution ratio of the _proc on the basis of the 5-line processed image signal S 5 reproduced image with respect to that of the reproduced on the basis of the base processed image signal S _proc B picture is in the parameter correcting means 306 (the frequency component will be referred to as the "target frequency component" hereinafter) (step S34) 306 corrects the parameters of the basic transformation functions f _kb on the basis of the filter characteristics F1 and F2 and the information about the target frequency component (step S35). Steps S4 and S5 may be performed prior to steps S2 and S3 or in parallel with them.

Then, the frequency emphasis processing means creates 303 band-limited signals on the basis of the unsharp image signals S _us k and carries out the frequency enhancement processing on the unsharp image signals S _us k using the order defined by the corrected parameter transformation functions, thereby 'to recover the intermediate processed image signal S _proc (step S36). Then, the enlargement / contraction device increases 305 the processed intermediate image signal S _proc 'to thereby obtain the final 5-line processed image signal S _pro c (step S37). Finally, an image is reproduced by means of a printer or the like on the basis of the 5-line processed image signal S _proc 5 (step S38).

Thus, in this embodiment, the parameters of the transform functions for performing the frequency emphasis on the original image signal S _org transformed in the pixel _{density are} set by correcting the parameters of the basic transform functions f _kb (the transform functions for performing the frequency emphasis of the basic original image signal S _org ') basis of the filtering characteristics F1 and F2 so that the frequency response of the processed image signal S _proc is substantially so as that of the base processed image signal S _proc B, and accordingly, the _proc on the basis of the processed image signal S, which is transformed from the picture element density of the original image signal S _{org was} obtained reproduced image in terms of sharpness equivalent to the image, which is reproduced on the basis of the processed basic image signal S _proc B, which was obtained from the basic original image _signal S _org '. In addition, since the frequency response is corrected simultaneously with the frequency emphasis processing, working time can be saved as compared with the case where the correction of the frequency response is performed with respect to the processed image signal S _proc which has been subjected to the frequency _emphasis processing.

In the seventh embodiment, although the parameters of the basic transform functions f _kb corresponding to the filter characteristic F1 of the pixel density transformation filter and the filter characteristic F2 of the interpolation filter are corrected, it is possible to prepare in advance a table in which the kind of the pixel density transformation means 304 The pixel density transformation to be performed and the type of the enlargement / contraction means to be performed 305 the interpolation filter to be used are related to the amount of correction of the parameters to correct the parameters with reference to the table based on the pixel density, the magnification, and the target frequency component input via the input means.

In addition, in the seventh embodiment, although the frequency response of the 5-line processed image signal S _proc 5 is based on the in 58 and 59 filter parameters F1 and F2 are obtained, and the parameters of the transformation function for transforming the band-limited image signal corresponding to the target frequency component are corrected so that the frequency response of the processed image signal S _proc from the 5-line transformed original image signal S _org 5 that of the processed basic image signal S _proc B is similar when the filter characteristics F1 and F2 are described by formulas, and the frequency response of the 5-line processed image signal S _proc 5 can be obtained from the formula.

Although in the seventh embodiment, the basic original image _signal S _org 'is transformed into the Original picture signal S _org transformed in the picture element _density , for which the basic original picture signal S _org 'is filtered by means of a picture element density transformation filter, the basic original picture signal S _org ' can also be converted into the original picture signal S _org transformed in the picture element density by other methods. For example, one can make use of linear interpolation or spline interpolation or thinning of the picture elements after filtering. In this case, the frequency response of the are in the picture element density-transformed original image signal S _org on the basis of the contraction coefficient of the interpolation calculated, or by means of a Fourier transform of the transformed original image signal element density in the image S _org, and the parameters of the base transformation functions f _kb corrected on the basis of the frequency response.

Is addition, when a processed image signal S _proc obtained by the fact that an intermediate processed image signal S _proc ', which consists of the base original image signal S _org' is obtained, as it is, by the expansion / contraction means 305 is increased, or when a processed image signal S _proc from the transformed in the pel density original image signal S _org without enlarging the intermediate processed image signal S _proc 'is obtained, the processed image signal S _proc can differ in frequency response characteristic from the base processed image signal S _proc B. In this case, the basic transformation functions can be corrected based on one of the filter characteristics F1 and F2.

Although in the seventh embodiment described above, the blur image signals obtained from the original image signal by filtering and interpolation / enlargement and the band-limited signals from the original image signal and the blur image signals can be won the band-limited signals also by, for example, transforming of the original image signal in multi-resolution image signals by wavelet transformation or a Laplace pyramid are generated, wherein the image signals with the respective resolutions be taken as bandlimited signals.

The following is an image processing system 331 explained according to an eighth embodiment of the invention.

In 65 contains the image processing system 331 according to the eighth embodiment of the invention, a blur image signal generating means 332 'Generated from a base original image signal S _org' having a plurality of unsharp image signals from multiple resolution image signals S _org 'are obtained by the transform into the multi-resolution space and encoding the transformed image signals, and a frequency enhancement processing means 333 That performs frequency enhancement processing for enhancing a particular frequency and an intermediate processed image signal S _proc 'is generated. The image processing system 331 also contains an enlargement / contraction device 335 processing the processed intermediate _{image signal} S _proc 'and a parameter correction means 336 , The from the blur image signal generator 332 , the frequency emphasis processing means 333 and the enlargement / contraction device 335 The processing performed is the same as in the seventh embodiment, and thus will not be described again here. In this embodiment, the parameter correcting means corrects 336 the parameters of the basic transform functions f _kb corresponding to the characteristics of the wavelet transform functions used when the original image signal S _{org is} transformed into the multi-resolution space.

The basic original image signal S _org is transformed into the multi-resolution space, and the transformed image signals are encoded in the following manner 66A is shown, the basic original image signal S _org 'is subjected to a wavelet transformation and decomposed into four pieces or blocks LL1, HL0, LH0 and HH0 by resolution. The data LL1 stands for an image obtained by reducing the original image to 1/2 in both the longitudinal and lateral directions. The data HL0, LH0 and HH0 correspond to a longitudinal edge, a side edge and an oblique edge. When the four data blocks LL1, LH0, LH0 and HH0 are subjected to inverse wavelet transformation, the original image signal S _{org is obtained} . Then, the data LL1 is subjected to another wavelet transformation and decomposed into four data blocks LL2, HL1, LH1 and HH1 as shown in FIG 66B is shown. The data LL2 represents an image obtained by reducing the data LL1 by 1/2 in both the longitudinal and lateral directions. The data HL1, LH1 and HH1 correspond to a longitudinal edge component, a side edge component and a skew edge component of the data LL1. When the four data blocks LL2, HL1, LH1 and HH1 are subjected to an inverse wavelet transformation, a low resolution image signal having a resolution 1/2 corresponding to that of the original image signal is obtained. Further wavelet transforms are performed in the desired frequency with respect to the data obtained each time a wavelet transform is performed, to thereby obtain a plurality of data blocks differing in resolution from each other. Subsequently, the data for each resolution according to 66C coded, and you get more Field-resolved image signals S _org ''.

By decoding only image signals to a desired resolution within the multi-resolution image signals S _org and performing the inverse wavelet transform with respect to these image signals, a low resolution image signal can be obtained, which is a 1/2 ^k resolution image Original image (k corresponds to the desired resolution).

As described above, by recovering the low-resolution image signals to a desired resolution from the multi-resolution image signals S _org '', a low resolution image signal for a 1/2 ^k resolution image (k corresponding to the desired resolution) of the Original image to be won. However, the image reproduced on the basis of the low-resolution image signal becomes blurred, depending on the wavelet transform functions used in the wavelet transform. 67 shows the frequency response of the image signals of low resolution, which are restored to a resolution of 1/2 of that of the original original image signal S _org 'and up to a resolution of 1/4 that of the original original image signal S _org ', if the latter by means of has been decomposed by wavelet transform functions having first filter coefficients, and 68 shows the same situation for the case that the base original image signal S _org 'by means of wavelet transformation functions of the second filter coefficients has been disassembled. Like from the 65 and 66 is removed, deteriorate with deteriorated response, the high-frequency components, and the image is less sharp. In the 65 and 66 is the frequency response of the basic original image signal S _org 'in the entire frequency bands 1 ,

The multi-resolution image signals S _org 'obtained by decomposing a basic original image signal S _org ' by wavelet transformation are combined with information about the size of the wavelet transformation functions and / or the associated filter coefficients, since the wavelet transformation functions used in the transformation should correspond to the one-to-one transform functions used in the inverse wavelet transform. Accordingly, in the eighth embodiment, a low resolution image signal having a desired resolution in the multiple resolution image signals S _org '' obtained by decomposing a base original image signal S _org ', recovered, and is taken as an object original image signal S _org, the difference is the response at a target frequency component between the processed image signal S _proc obtained by performing frequency emphasis on the object original image signal S _org using the basic transform functions f _kb and the processed basic image signal S _proc B obtained by performing the frequency _emphasis processing on the fundamental original image signal S _org 'by use of the base transformation functions f _kb, obtained on the basis of the information on the wavelet transformation functions which the multiple resolution image signals S _org' match ', and the parameters of the transformation functions for transformants ieren of the object original image signal S _org be from the parameter correction means 336 corrected so that the difference is compensated. If information about the wavelet transformation functions is not added to the multi-resolution image signals S _org '', then the information is manually transferred to the image processing system 331 entered.

For example, when a reproduced image signal of a low resolution than an object original image signal S _{org is} taken up to a resolution of 1/4 that of the original original image signal S _org obtained by decomposing the basic original image signal S _org ', the resolution ratio ( Aspect) of the image reproduced on the basis of the 1/4 resolution image signal at a frequency of 1 cycle / mm, adapted in the following manner to that of the image reproduced on the basis of the original original image signal S _org ': the frequency at which the resolution ratio of the image reproduced on the basis of the 1/4 resolution image signal is to be matched with that of the image reproduced on the basis of the original original image signal S _org ', is input in advance by the user via an input device. Since the response of the image reproduced on the basis of the 1/4 resolution image signal at 1 cycle / mm is 0.9, while the response of the image reproduced on the basis of the original original image signal S _org 'at 1 cycle / mm is 1, it is necessary to multiply the frequency response at 1 cycle / mm of the image with a resolution of 1/4 by 1 / 0.9 to equalize the resolution ratio of 1/4 resolution based on the image to that of the reason Original image signal S _org 'with 1 cycle / mm reproduced image. Consequently, in this embodiment, the band limited signal transforming function in the 1 cycle / mm corresponding frequency band is adjusted by correcting the parameters for the fundamental transform function f _kb for the band limited signal in the frequency band corresponding to 1 cycle / mm so that the Extent of enlargement equal to 1 / 0.9 times. In particular, if the transformation function is a function with a slope, the slope is multiplied by 1 / 0.9, and if the transformation function is a constant, it is multiplied by 1 / 0.9. The frequency component in which the resolution ratio of the image reproduced from the 1/4 resolution image signal is the same That is, the image reproduced from the original original image signal S _org 'may be any frequency, but may include two or more frequencies. Furthermore, the frequency component may be input from the user via an input device, not shown.

If in this embodiment the frequency enhancement processing is carried out of an image signal with respect to which a lower resolution than a base original image signal S _org ', and from multiple resolution image signals S _org' 'is obtained, which in turn by wavelet transformation of the base original image signal S _org are obtained, the transform functions for transforming the band-limited signals obtained from the low resolution image signal are corrected on the basis of the wavelet transform functions used in the wavelet transform of the basic original image signal S _org , so that based on the processed image signal from the image signal of low resolution reproduced image substantially equal to the image which is reproduced on the basis of the processed image signal obtained from the original image signal, with respect to the frequency response at a desired frequency. Consequently, a processed picture signal whose frequency response is substantially equal to that of a processed picture signal can be constantly obtained from the basic original picture signal, irrespective of the resolution of the low-resolution picture signal. In addition, since the frequency response is corrected simultaneously with the frequency emphasis processing, working time can be saved as compared with a correction of the frequency response made with respect to the processed image signal S _proc which has been subjected to the frequency emphasis.

If the processed image signal S _proc from the expansion / contraction device 335 In the eighth embodiment, the parameters of the basic transformation functions f _kb are calculated on the basis of the filter characteristic of the magnification / contraction means 335 corrected in addition to the characteristics of the wavelet transform functions.

In the seventh and eighth embodiments described above, it is preferable that the object original image _signal S _org and the parameters for the object original image _signal S _org obtained by correcting the basic transformation functions correspond to the characteristic of the magnification or the characteristic of the contraction, correlated with each other. In this embodiment, when an image is outputted (output as a hard copy on a photographic film or displayed on a screen or the like), again based on the original object image signal S _org , an image having a quality similar to that of a previous one can be obtained output image resembles.

Further, it is preferable that the basic original image _signal S _org 'from which the object original image _signal S _{org is} obtained, the magnification characteristics or those for the contraction with which the basic original image _signal S _org ' is increased or contracted the object original image signal S _org , and the parameters of the reference transformation functions are correlated with each other. This measure enlarged or reduced images can be reproduced repeatedly on the basis of the base original image signal S _org 'with the same image quality.

The characteristic of the magnification or the characteristic of the contraction can be represented by a spatial frequency characteristic and stored in the form of tabular data, as shown in FIG 69 are shown. Otherwise, the spatial frequency characteristics may be approximated by a function according to the characteristic of the magnification or the characteristic of the contraction, and parameters of this function may be stored as an enlargement or contraction characteristic. As such a function, a Gaussian function of the following kind can be used: Res = (-a 2 / σ 2 ) where a = 0, if Fq - m <0, a = Fq - m otherwise; Res the answer and Fq means the spatial frequency. In this case, σ and m may be used as the parameters corresponding to the space frequency characteristic, which in turn represents the characteristic of the magnification or the characteristic of the contraction, as in FIGS 70A and 70B is shown.

Claims (102)

An image processing method of obtaining a processed image signal from an original image signal representing an original image with a certain resolution, wherein a plurality of frequency-band-limited signals are generated based on the original image signal, obtaining a plurality of transformed image signals by performing a transform processing on the bandbe bordered signals are performed on the basis of associated transformation functions, and a processed image signal in which a predetermined frequency component is emphasized is obtained from the transformed image signals, characterized by the step of using predetermined transformation functions corresponding to the respective bandlimited signals on the basis of Resolution of the original image signal. The method of claim 1, wherein the transformation functions are defined by transformation function definition parameters for the Transformation functions based on the resolution of the original image be determined. The method of claim 1, wherein the plurality of band-limited ones Signals are formed by that with respect to the original image signal a filter processing is performed using filters whose Filter coefficients are determined based on the pixel density the original image signal to thereby form a plurality of blur image signals to create, which differ in their frequency response, and um a plurality of band-limited signals representing the signals in the individual frequency bands represent the original image signal, based on the blur image signals and the original image signal. The method of claim 1, wherein the predetermined Transformation functions are non-linear functions. The method of claim 1, wherein the transformation processing is a frequency emphasis processing. The method of claim 1, wherein the transformation processing is a dynamic range compression processing. The method of claim 1, comprising the step storing parameters defining the transformation function, the for the original image signal is determined based on the original image signal Parameters. The method of claim 1, wherein the step of Defining the transformation functions by determining the Transformation function defining parameters for the transformation functions based on the resolution of the original image contains the following steps: Prepare Transformation Function Definition Parameters for at least two reference resolutions, Compare the resolution of the original image with the reference resolutions and determining the transformation function definition parameters for one the reference resolutions, the resolution closest to the original image comes, as a transformation function definition parameter for the original image signal. Image processing system for obtaining a processed Image signal from an original image representing a certain resolution Original image signal comprising a band-limited signal generator, the multiple band limited based on the original image signal Generates signals that differ from one another in the frequency band, and a transformation device that transformed several Image signals thereby wins that a Transformation processing with respect to the band-limited signals based on associated transformation functions carried out is to a processed image signal in which a predetermined Frequency component is highlighted from the transformed image signals to win, characterized by a transformation function defining means, the predetermined transformation functions corresponding to the respective band-limited ones Signals, based on the resolution of the original image signal used. The system of claim 9, wherein the transformation function defining means Defines the transformation functions by using transformation function definition parameters for the Transformation functions based on the resolution of the Original image determined. The system of claim 10, wherein the band limited signal generator a blur image signal generator contains the multiple blur image signals forms, which have a different frequency response, in the in terms of of the original image signal, filter processing using Filtering performed whose filter coefficients are based on the resolution of the Original image are determined, and the band-limited signal generating means as the band-limited signals creates multiple signals, which the signals in the individual frequency bands of the original image signal represent, based on the blur image signals and the original image signal. The system of claim 9, wherein the predetermined ones Transformation functions are non-linear functions. The system of claim 9, wherein the transformation processing is a frequency emphasis processing. The system of claim 9, wherein the transformation processing is a dynamic range compression processing. The system of claim 9, further comprising Means for storing the transformation function definition parameters, the for the original image signal with respect to the original image signal Parameters are determined. A system according to claim 9, wherein said transformation function determining means in itself transformation function definition parameter for at least two reference resolutions stores, the resolution of the original image corresponding to the original image signal with the reference resolutions and the transformation function definition parameters for the one Reference resolution, the resolution closest to the original image comes when the transformation function definition parameter determines for the original image signal. Computer readable recording medium with one on top stored program for causing a computer, an image processing method for obtaining a processed image signal from an original image with a certain resolution representing Perform original image signal, in which several band-limited signals that are in the frequency band from each other be differentiated based on the original image, a plurality of transformed image signals are obtained by a transformation processing in terms of the individual band-limited signals on the basis of associated transformation functions carried out becomes, and from the transformed image signals a processed Image signal is obtained, in which a predetermined frequency component is highlighted, characterized in that the program is the step involves using predetermined transformation functions, which the individual bandlimited signals on the basis of resolution correspond to the original image signal. A recording medium according to claim 17, wherein said Transformation functions are defined by determining transformation function definition parameters for the Transformation functions based on the resolution of the Original image. A recording medium according to claim 18, wherein said Step of Generating Multiple Bandlimited Signals Following Steps has: perform a filter processing with respect to of the original image signal using filters whose filter coefficients based on the resolution of the original image to thereby obtain a plurality of blur image signals to form, which differ in the frequency response, and by several band-limited signals, which signals in the individual Frequency bands of the Represent original image signal, based on the blur image signals and the original image signal. A recording medium according to claim 18, wherein said predetermined transformation functions non-linear functions are. A recording medium according to claim 18, wherein said Transformation processing a frequency emphasis processing is. A recording medium according to claim 18, wherein said Transformation processing a dynamic range compression processing is. A recording medium according to claim 18, wherein said The program further includes the step of storing the transformation function definition parameters that includes for the original image signal determined with respect to this parameters become. The recording medium of claim 18, wherein the step of defining the transformation functions by determining transformation function definition parameters for the transformation functions based on the resolution of the original image comprises the steps of: preparing Transformation function definition parameters for at least two reference resolutions, comparing the resolution of the original image corresponding to the original image signal with the reference resolutions, and determining the transformation function definition parameters for the reference resolution closest to the resolution of the original image signal as the transformation function definition parameter for the original image signal. The method of claim 1, wherein the transformation functions are determined by from the resolution of the object original image signal dependent Correct reference transformation functions for transformation from a reference original image signal containing an image with represents a reference resolution, obtained band-limited signals, wherein the reference transformation functions determined and saved in advance. The method of claim 25, wherein the plurality band-limited signals are generated by the fact that respect Original image signal filter processing using Filtering is performed with predetermined filter coefficients, thereby several blur image signals to generate, which differ in the frequency response, and a plurality of band-limited signals are formed, which signals in the individual frequency bands represent the original image signal, based on the blur image signals and the original image signal. An image processing method according to claim 25, wherein the transformation functions are non-linear functions. The method of claim 25, wherein the transformation processing is a frequency emphasis processing. The method of claim 25, wherein the transformation processing is a dynamic range compression processing. An image processing method according to claim 25, wherein the predetermined transformation functions with the on the Original image signal related functions are stored. The method of claim 25, in which are calculated the transformation functions for band-limited signals for the to be processed original image signal such that the frequency response of the processed image signal S _proc to that of the reference processed image signal S _proc 0 at least equal in frequency bands not lower than 1/5 of the Nyquist frequency of the original image signal to be processed. A method according to claim 28, wherein the band limited signal transforming functions for the original picture signal to be processed are calculated such that the frequency response of the processed picture signal S _proc corresponds to that of the processed reference picture signal S _proc 0 at least in frequency bands not lower than 1/2 Nyquist frequency of the original image signal to be processed. The system of claim 9, wherein the transformation function defining means includes a transformation function calculating means which the transformation function to process from a to be processed Object original image signal obtained band-limited signals thereby calculated that she dependent from the resolution of the original object image signal reference transform functions for Transformed the intermediate image signals corrected from a Reference original image signal obtained with a reference resolution with the reference transformation functions determined in advance and saved. A system according to claim 33, wherein said bandlimited signal generator means comprising: a blur image signal generator, the multiple blur image signals forms, which differ in the frequency response by in terms of of the original image signal using filter processing performed by filters whose filter coefficients are based on the resolution of the Original image signal are determined, and as band-limited signals generates a plurality of band-limited signals which the signals in the individual frequency bands represent the original image signal, based on the blur image signals and the original image signal. The system of claim 33, wherein the predetermined ones Transformation functions are non-linear functions. The system of claim 33, wherein the transformation processing is a frequency emphasis processing. The system of claim 33, wherein the transformation processing is a dynamic range compression processing. The system of claim 33, further comprising Device for storing the predetermined transformation functions with the functions related to the original picture signal. A system according to claim 33, wherein said transformation function calculating means the transformation functions to process from one to be processed Object original image signal obtained band-limited signals thereby calculated that she dependent from the resolution of the object original image signal Reference transformation functions corrected so that the frequency response the processed image signal with that of the processed reference image signal at least in frequency bands, which are not lower than 1/5 of the Nyquist frequency of the processed one Original image signal. System according to claim 39, in which the transformation functions for band-limited signals are calculated for the material to be processed original image signal such that the frequency response of the processed image signal S _proc to that _proc 0 at least equal in frequency bands of the reference processed image signal S, which are not lower than 1/2 of the Nyquist frequency of the original image signal to be processed. Computer-readable recording medium according to claim 17, in which the program is the step of calculating the transformation functions for processing the object original image signal to be processed includes sustained band limited signals, which performed by dependent on the resolution of the object original image signal correction of reference transformation functions for transforming the from a reference original image signal with a reference resolution obtained intermediate image signals, wherein the reference transformation functions determined and saved in advance. A recording medium according to claim 41, wherein said Step of generating the plurality of band-limited signals following Steps includes: Performing a filter processing with respect to of the original image signal using filters whose filter coefficients be determined on the basis of the resolution of the original image signal, thereby several blur image signals to obtain, which differ in the frequency response, and in that a plurality of band-limited signals representing the signals in the individual frequency bands represent the original image signal, are used as intermediate image signals based on the blur image signals and the original image signal. A recording medium according to claim 41, wherein said predetermined transformation functions non-linear functions are. A recording medium according to claim 41, wherein said Transformation processing a frequency emphasis processing is. A recording medium according to claim 41, wherein said Transformation processing a dynamic range compression processing is. A recording medium according to claim 41, wherein said Program continues the step of saving the specific transformation functions includes functions related to the original image signal. A recording medium according to claim 41, wherein said Reference transformation functions depending on the resolution of the Object original image signal to be corrected so that the frequency response of the processed image signal to that of the processed reference image signal at least in frequency bands which are not lower than 1/5 of the Nyquist frequency of the original image signal to be processed. A recording medium according to claim 47, wherein said Reference transformation functions depending on the resolution of the Object original image signal to be corrected so that the frequency response of the processed image signal to that of the processed reference image signal at least in frequency bands which are not lower than 1/2 of the Nyquist frequency of the original image signal to be processed. Method according to claim 1, comprising the following steps: Creating reference transformation functions for transforming the band-limited signals obtained from a reference original image signal representing a reference original image with a reference resolution set according to the frequency bands of the associated band-limited signals, and when an object-original image signal to be processed for an original image having a resolution lower than the reference resolution, setting the transformation functions for transforming the band-limited signals obtained from the object original image signal to be equal to the respective reference transform functions for the band-limited signals in the frequency bands that are not are higher than the frequency band corresponding to the resolution of the original image represented by the object original image signal. The method of claim 49, wherein the plurality Bandlimited signals are created by the fact that on the Basis of the original image signal, a plurality of blur image signals are generated which differ in frequency response, and several band limited signals for the signals in the individual frequency bands of the original image signal based on the blur image signals and the original image signal. The method of claim 49, wherein information about the resolution of the object original image signal, and the step of Setting transformation functions to transform the Intermediate image signals obtained from the object original image signal will be executed on the basis of that information. The system of claim 9, wherein the transformation processing means Reference transform functions to transform band-limited Generates signals obtained from a reference original image signal which is for is a reference original image with a reference resolution, which accordingly the frequency bands the individual band-limited signals are set, and, if the object original image signal to be processed is an original image represents its resolution less than the reference resolution is the transformation functions to transform the from the Object original image signal obtained band-limited signals so adjusts that she the associated Reference transformation functions for the band-limited signals in the frequency bands same, not higher lie as the frequency band corresponding to the resolution of the through the object original image signal represented Original image. The system of claim 52, wherein said bandlimited signal generator comprising: a blur image signal generator, the several based on the original image in the frequency response different blur image signals generates and as intermediate image signals a plurality of band-limited signals generates the signals in the individual frequency bands of the Represent original image, based on the blur image signals and the original image signal. A system according to claim 52, wherein a resolution information obtaining means for gaining information about the resolution of the original image represented by the object original image signal is, and the transformation processing means the transformation functions for transforming the intermediate image signals obtained from the object original image signal on the basis of this information. A recording medium according to claim 17, wherein said Program includes the following steps: Creating reference transformation functions for transforming the band-limited signals resulting from a reference original image signal were received, which for a reference original image with a reference resolution that depends on the frequency bands the individual band-limited signals are set, and, if the object original image signal to be processed for an original image with a resolution is less than the reference resolution, the transformation functions for transforming the band-limited signals from the object original image signal adjust so that they are equal to the reference transform functions for the band-limited ones Signals in the frequency bands, not higher lie as the frequency band corresponding to the resolution of the original image represented by the object original image signal. A recording medium according to claim 55, wherein said a plurality of band-limited signals are produced by the fact that on the Basis of the original image signal, a plurality of blur image signals are generated which differ in frequency response, and several band-limited signals are generated, represented by the signals in the individual frequency bands of the original image signal based on the blur image signals and the original image signal. A recording medium according to claim 55, wherein information about the resolution of the object origi and the step of setting the transformation functions for transforming the band-limited signals from the object original image signal is performed on the basis of that information. A method according to claim 1 for generating a contraction-processed image signal representative of an image having a desired resolution lower than that of an image represented by a processed reference image signal, the processed reference image signal being generated by selecting from a reference image for a reference original image. Original image signal, a first to nth reference image of low resolution are generated, wherein the reference image signals differ from each other in the frequency band and each represent images whose resolutions are 1/2 ^k (k = 1 to n) of the resolution of the reference original image signal; based on the low-resolution reference picture signals, low-resolution band-limited reference signals are generated; obtaining a plurality of transformed reference image signals by performing a predetermined transformation processing on the low-resolution band-limited reference signals on the basis of a plurality of reference transformation functions which are set depending on the respective frequency bands of the low-resolution band-limited reference signals; and predetermined processing is performed on the transformed reference image signals, wherein the image processing method is characterized by the steps of: taking as an original image signal one of the first to n-th low resolution reference image signals representing an image whose resolution of the desired resolution is determined by Reference image signals of low resolution represent images closest to each other; from the original image signal, first to m-th low-resolution image signals are generated which are different from each other in the frequency band and represent images whose resolutions are 1/2 ^k (k '= 1 to m) of the resolution of the original image signal; based on the low resolution image signals, low resolution band limited signals are generated; obtaining a plurality of transformed image signals by performing the predetermined transformation processing on the low-resolution band-limited signals based on a part of the reference transformation functions set for the respective frequency bands of the low resolution band-limited reference signals not higher than the frequency band; which corresponds to the desired resolution closest to resolution; the predetermined processing of the transformed image signals is performed to thereby obtain a processed intermediate image signal; the processed intermediate image is enlarged or contracted, according to which the resolution of the image represented by the processed intermediate image signal equals the desired resolution; and the enlarged or contracted processed intermediate image signal is used as a contraction processed image signal. The method of claim 58, wherein the band-limited ones Low resolution signals be generated by that on the basis of the original image signal a plurality of blur image signals be generated, which differ in the frequency response, and a plurality of band-limited signals are generated for the signals in the individual frequency bands of the original image signal based on the blur image signals and the original image signal. An image processing system according to claim 9, configured to generate a contraction-processed image signal representative of an image having a desired resolution lower than that of an image represented by a processed reference image signal, the processed reference image signal being generated from a reference original image Reference original image signal to a first to nth reference image low resolution are generated, the reference image signals differ from each other in the frequency band and each represent images whose resolutions are 1/2 ^k (k = 1 to n) of the resolution of the reference original image signal ; based on the low-resolution reference picture signals, low-resolution band-limited reference signals are generated; obtaining a plurality of transformed reference image signals by performing a predetermined transformation processing on the low-resolution band-limited reference signals on the basis of a plurality of reference transformation functions which are set depending on the respective frequency bands of the low-resolution band-limited reference signals; and predetermined processing is performed on the transformed reference image signals, the image processing method characterized by the steps of selecting means selecting as original image signal one of the first to n-th low resolution reference image signals representing an image whose resolution is the desired resolution in comes closest to the images represented by the low resolution reference image signals; a low-resolution image signal generating means which generates first to m-th low-resolution image signals from the original image signal which are different from each other in their frequency band and which represent images whose resolutions are 1/2 ^{k '} (k' = 1 to m) of Resolution of the original image signal amount; a bandbe low-resolution signal limited-signal generating means for generating low-resolution band-limited signals on the basis of the low-resolution image signals; transformation processing means for obtaining a plurality of transformed image signals by performing the predetermined transformation processing on the low-resolution band-limited signals on the basis of a part of the reference transformation functions set for the individual frequency bands of the low-resolution reference image signals which are not higher as the frequency band corresponding to the resolution that comes closest to the desired resolution, and the predetermined processing with respect to the transformed image signals, thereby obtaining a processed intermediate image signal; and an enlargement / contraction means which enlarges or contracts the processed intermediate image such that the resolution of the image represented by the processed intermediate image signal equals the desired resolution. The system of claim 60, wherein said bandlimited signal generator based on the original image signal, a plurality of blur image signals generated, which differ in the frequency response, and based on the blur image signals and producing a plurality of band-limited signals from the original image signal, the signals in the individual frequency bands of the original image signal represent. The computer-readable recording medium of claim 17, wherein the program causes a computer to perform a method of generating a contraction-processed image signal representing an image, wherein the processed reference image signal is generated by converting first to nth reference original image signal from a reference original image signal the low-resolution reference image signals differ from each other in the frequency band and each represent images whose resolutions are 1/2 ^k (k = 1 to n) of the resolution of the reference original image signal; based on the low-resolution reference picture signals, low-resolution band-limited reference signals are generated; obtaining a plurality of transformed reference image signals by performing a predetermined transformation processing on the low-resolution band-limited reference signals on the basis of a plurality of reference transformation functions which are set depending on the respective frequency bands of the low-resolution band-limited reference signals; and predetermined processing is performed on the transformed reference image signals, the program including the steps of: taking as an original image signal one of the first to n-th low resolution reference image signals representing an image whose resolution of the desired resolution is smaller than that of the reference image signals Resolution represented images comes closest; from the original image signal, first to m-th low-resolution image signals are generated which are different from each other in the frequency band and represent images whose resolutions are 1/2 ^{k '} (k' = 1 to m) of the resolution of the original image signal; based on the low resolution image signals, low resolution band limited signals are generated; obtaining a plurality of transformed image signals by performing the predetermined transformation processing on the low-resolution band-limited signals based on a part of the reference transformation functions set for the respective frequency bands of the low resolution band-limited reference signals not higher than the frequency band; which corresponds to the desired resolution closest to resolution; the predetermined processing of the transformed image signals is performed to thereby obtain a processed intermediate image signal; the processed intermediate image is enlarged or contracted, according to which the resolution of the image represented by the processed intermediate image signal equals the desired resolution; and the enlarged or contracted processed intermediate image signal is used as a contraction processed image signal. A recording medium according to claim 62, wherein said band-limited signals of low resolution are generated thereby that on the basis of the original image signal a plurality of blur image signals be generated, which differ in their frequency response, and that several band limited signals representing the signals in the individual frequency bands of the Represent original image signal, based on the blur image signals and the original image signal. The method of claim 1, comprising the steps of: preparing reference transformation functions for transforming the band-limited signals obtained from a reference original image signal for a reference original image with a reference resolution set in accordance with the frequency bands of the individual band-limited signals, and when an object-to-be-processed image signal represents an original image having a resolution higher than the reference resolution, setting the transform functions to transform the band-limited signals from the object original image signal into frequency bands not higher are as the frequency band corresponding to the reference resolution so as to resemble the respective reference transformation functions, and the transformation functions for transforming the band-limited signals from the object original image signal into frequency bands higher than the frequency band corresponding to the reference resolution so as to be the reference transformation function for transforming of the band-limited signal is equal to that obtained from the reference original image signal in the highest frequency band. The method of claim 64, wherein the plurality band limited signals are generated by based on the Original image signal multiple blur image signals are generated which differ in the frequency response from each other, and that on the Basis of the blur image signals and the original image signal generates a plurality of band-limited signals which are representative are for the signals in the respective frequency bands of the original image signal. The method of claim 64, wherein information about the resolution of the object original image signal, and in which the step setting the transformation functions to transform the intermediate image signals obtained from the object original image signal will be executed on the basis of that information. The method of claim 64, further comprising the step of storing the transformation function definition parameters, the for the original image signal was determined with the related image signal Parameters. An image processing system according to claim 9, wherein a transformation processing device reference transformation functions for transforming the band-limited signals prepared by a reference original image signal obtained for a reference original image with a reference resolution stands, the transformations depending on the frequency bands of individual band-limited signals are set, and, if one to be processed object original image signal for an original image with a resolution stands taller than the reference resolution is they use the transformation functions to transform the the band-limited signals obtained from the original object image signal in frequency bands that not higher are as the reference resolution corresponding frequency band, so that they are the relevant reference transformation functions same, and the transformation functions to transform the band-limited signals received from the object original image signal in frequency bands above the reference resolution corresponding frequency bands are so adjusted that they equal the reference transform function to transform the band-limited Are signals which are from the reference original image signal in the highest frequency band is obtained. The system of claim 68, wherein said band limited signal generator a blur image signal generator which has a plurality of blur image signals generated, which differ in their frequency response, based on the original image, and multiple band-limited signals which generates the signals in the respective frequency bands of the Represent original image signal based on the blur image signals and the original image signal. A system according to claim 68, wherein a resolution information extraction means to get information about the resolution of the original image represented by the object original image signal is, and the transformation processing means the transformation functions for transforming the intermediate image signals received from the object original image signal on the basis of that information. The system of claim 68, further comprising Means for storing the transformation function definition parameters, the for the original image signal was determined with the related image signal Parameters. The computer-readable recording medium of claim 17, wherein the program comprises the steps of: preparing reference transformation functions for transforming the band-limited signals obtained from a reference original image signal for a reference original image at a reference resolution corresponding to the frequency bands of the individual band-limited ones Signals are set, and when an object-to-be-processed original image signal represents an original image having a resolution higher than the reference resolution, setting the transform functions to transform the band-limited signals from the object original image signal into frequency bands not higher are as the frequency band corresponding to the reference resolution so as to resemble the respective reference transformation functions, and the transformation functions for transforming the band-limited signals from the object original image signal into frequency bands higher than the frequency band corresponding to the reference resolution so as to be the reference transformation function for transforming of the band-limited signal is equal to that obtained from the reference original image signal in the highest frequency band. A recording medium according to claim 72, wherein said Several band-limited signals are generated based on the original image signal, a plurality of blur image signals are generated, which differ in the frequency response from each other, and that on the Basis of the blur image signals and the original image signal generates a plurality of band-limited signals which are representative are for the signals in the respective frequency bands of the original image signal. A recording medium according to claim 72, wherein information about the resolution of the object original image signal, and in which the step setting the transformation functions to transform the intermediate image signals obtained from the object original image signal will be executed on the basis of that information. A recording medium according to claim 72, wherein said Program further includes the step of storing the transformation function definition parameters, the for the original image signal was determined with the related image signal Includes parameters. The method of claim 1, wherein, when an processing object original image signal is a contracted image signal obtained by executing a Resolution transformation processing in terms of a reference original image signal representing an image with a reference resolution, the transformation functions for processing the object original image signal obtained band limited signals are determined by depending on the characteristic of the resolution transformation processing Reference transformation functions to transform the from Reference original image obtained intermediate image signals are corrected such that the frequency response a desired one Frequency component of the processed received from the object original image signal Image signal equivalent becomes that of the processed image signal, that of the reference original image signal was obtained, with the reference transformation functions in advance were determined and stored. The method of claim 76, wherein when said processed image signal obtained from the contracted image signal magnified with an enlargement, the transformation functions for processing the object original image signal obtained band-limited signals are determined by the reference transformation functions according to both the Characteristic of the resolution transformation processing as well as the characteristics of the magnification. The method of claim 76, wherein the plurality Bandlimited signals are generated by the original image signal several blur image signals be generated, which differ in the frequency response, and based on the blur signals and the original image signal generates a plurality of band-limited signals which are the signals in the respective frequency bands of the Represent original image signal. The method of claim 76, wherein the object original image signal and the parameters for the object original image signal obtained by correcting the reference transformation functions according to the contraction characteristic was obtained, correlated with each other. The method of claim 76, wherein the reference original image signal, from which the object original image signal is obtained, the contraction characteristic, with which the reference original image signal is contracted, and the parameters of the reference transformation functions correlate be stored together. The method of claim 1, wherein, when a enlargement processing subjected image signal by enlarging the processed image signal with a desired Obtained magnification should be, by a magnification transformation is performed on the processed image signal, the transformation functions for transforming the respective band-limited signals depending on the characteristic of magnification transformation be corrected so that the Frequency response of a desired Frequency component of the enlarged processed Image signal to a predetermined frequency response is. The method of claim 81, wherein the plurality Bandlimited signals are generated by the original image signal several blur image signals be generated, which differ in the frequency response, and that based on the blur image signals and the original image signal generates a plurality of band-limited signals which are the signals in the respective frequency bands of the Represent original image signal. The method of claim 81, wherein the object original image signal and the parameters for the object original image signal obtained by correcting the reference transform functions according to the magnification characteristic be stored together. The method of claim 81, wherein the reference original image signal, from which the object original image signal is obtained, the magnification characteristic, with the reference original image signal to the object original image signal is enlarged, and the parameters of the reference transformation functions correlate be stored together. An image processing system according to claim 9, comprising: a Transformation function correction device, which, if one too processing object original image signal, a contracted image signal is gained by running a resolution transformation processing in terms of a reference original image signal representing an image with a reference resolution, the transformation functions for processing the object original image signal obtained band-limited signals determined by that depends on the characteristic of the resolution transformation processing Reference transformation functions to transform the from Reference original image obtained band-limited signals so determined that the Frequency response of a desired Frequency component of the obtained from the object original image signal processed image signal equivalent becomes the one of the processed image signal, which is from the Reference original image signal was obtained, using the reference transformation functions are determined in advance and stored. A system according to claim 85, wherein, if that of the contracted image signal received processed image signal with a Magnification is increased, the transformation function correcting means the transformation functions for processing the image obtained from the object original image signal Intermediate image signals determined by passing the reference transformation functions according to both the characteristic of the resolution transformation processing as well as the magnification characteristic corrected. The system of claim 85, wherein the plurality of band-limited ones Image signals are generated by that from the original image signal several blur image signals be generated, which differ in the frequency response, and that based on the blur image signals and the original image signal generates a plurality of band-limited signals which are the signals in the individual fringe bands of the Represent original image signal. The system of claim 85, further comprising Means for storing the object original image signal and the Parameters for the object original image signal, which by correcting the reference transformation functions were obtained according to the contraction characteristic, in correlated way. The system of claim 85, further comprising Device for storing the reference original image signal, from which the object original image signal was obtained, the contraction characteristic, with which the reference original image signal to the object original image signal is contracted, and the parameter of the reference transformation functions in a correlated manner. The system of claim 9, comprising: a transformation function correcting device, which, if one of a magnification processing subjected image signal by enlarging the processed image signal with a desired Magnification can be obtained should, by a magnification transformation in terms of the processed image signal is performed, the transformation functions for transforming the individual band limited signals depending on the Characteristic of the magnification transformation corrected so that the Frequency response of a desired Frequency component of the subjected to a magnification processing Image signal to a predetermined frequency response is. The system of claim 90, wherein the plurality of band-limited signals are generated by generating from the original image signal a plurality of blur image signals differing in frequency response from differ, and that based on the blur image signals and the original image signal, a plurality of band-limited signals are generated which represent the signals in the individual frequency bands of the original image signal. The system of claim 90, further comprising Means for storing the object original image signal and the Parameters for the object original image signal, by correspondingly correcting the reference transformation functions the magnification characteristic obtained in a correlated manner. The system of claim 90, further comprising Device for storing the reference original image signal, from which the original original image signal is obtained, the magnification characteristic, with the reference original image signal on the object original image signal is enlarged, and the parameter of the reference transformation functions in each other correlated way. A recording medium according to claim 17, wherein said Program includes the step that, if one to be processed Object original image signal is a contracted image signal, won by running a resolution transformation processing in terms of a reference original image signal representing an image with a reference resolution, the transformation functions for processing the object original image signal obtained band limited signals are determined by depending on the characteristic of the resolution transformation processing Reference transformation functions to transform the from Reference original image obtained intermediate image signals are corrected such that the frequency response a desired one Frequency component of the processed received from the object original image signal Image signal equivalent becomes that of the processed image signal that is from the reference original image signal was obtained, with the reference transformation functions in advance were determined and stored. A recording medium according to claim 94, wherein said Program further includes the step that, if that of the contracted Image signal obtained processed image signal is magnified by a magnification, the transformation functions for processing the object original image signal obtained band-limited signals are determined by the reference transformation functions according to both the characteristic of the resolution transformation processing as well as the characteristics of the magnification. A recording medium according to claim 94, wherein said several band-limited signals are generated by the original image signal several blur image signals be generated, which differ in the frequency response, and based on the blur signals and the original image signal generates a plurality of band-limited signals which are the signals in the respective frequency bands of the original image signal represent. A recording medium according to claim 94, wherein said Object original image signal and the parameters for the object original image signal, obtained by correcting the reference transformation functions according to the contraction characteristic was correlated with each other. A recording medium according to claim 94, wherein said Program further comprises the step of storing the reference original image signal, from which the object original image signal is obtained, the contraction characteristic, with the reference original image signal on the object original image signal is contracted, and the parameter of the reference transformation functions in a correlated manner. A recording medium according to claim 17, wherein said Program includes the step in which, if an enlargement processing subjected image signal by enlarging the processed image signal with a desired Obtained magnification is to be done by a magnification transformation the processed image signal is performed, the transformation functions for transforming the respective band-limited signals depending on the characteristic of magnification transformation be corrected so that the Frequency response of a desired Frequency component of the enlarged processed Image signal to a predetermined frequency response is. A recording medium according to claim 99, wherein the plurality of band-limited signals are generated by generating a plurality of blur image signals different in frequency response from the original image signal, and generating a plurality of band-limited signals based on the blur image signals and the original image signal respective frequency bands represent the original picture signal. A recording medium according to claim 99, wherein said Object original image signal and the parameters for the object original image signal, which are obtained by correcting the reference transformation functions according to the magnification characteristic, correlated with each other. A recording medium according to claim 99, wherein said Program further includes the step of the original reference picture signal, from which the object original image signal is obtained, the magnification characteristic, with the reference original image signal on the object original image signal is enlarged, and the parameters of the reference transformation functions in each other be stored in a correlated manner.